paper_id stringlengths 10 10 | year int64 2.02k 2.03k | conference stringclasses 3
values | accepted bool 2
classes | title stringlengths 14 183 | abstract stringlengths 246 3.6k | keywords listlengths 1 28 | pdf_url stringlengths 40 40 | paper_text stringlengths 53 566k ⌀ | reviews listlengths 2 8 |
|---|---|---|---|---|---|---|---|---|---|
3aVZhMfsyz | 2,023 | NeurIPS 2023 | true | Volume Feature Rendering for Fast Neural Radiance Field Reconstruction | Neural radiance fields (NeRFs) are able to synthesize realistic novel views from multi-view images captured from distinct positions and perspectives. In NeRF's rendering pipeline, neural networks are used to represent a scene independently or transform queried learnable feature vector of a point to the expected color o... | [
"neural rendering",
"volume rendering",
"view synthesis",
"3D reconstruction"
] | https://openreview.net/pdf?id=3aVZhMfsyz | Volume Feature Rendering for Fast Neural Radiance Field Reconstruction
Abstract
Neural radiance fields (NeRFs) are able to synthesize realistic novel views from multi-view images captured from distinct positions and perspectives. In NeRF's rendering pipeline, neural networks are used to represent a scene independentl... | [
"{\"IS_META_REVIEW\": false, \"comments\": \"SUMMARY:\\nThis work improves grid-based NeRF on both quality and training speed. To predict view-dependent color, they proposes to condition MLP on the volume rendered voxel features instead of each original point features. As a result, the MLP only run once for each pi... |
3b5e2AFs7f | 2,023 | NeurIPS 2023 | false | On Formal Feature Attribution and Its Approximation | Recent years have witnessed the widespread use of artificial intelligence (AI) algorithms and machine learning (ML) models. Despite their tremendous success, a number of vital problems like ML model brittleness, their fairness, and the lack of interpretability warrant the need for the active developments in explainable... | [
"Feature Attribution",
"Explainable AI",
"Formal Explanation"
] | https://openreview.net/pdf?id=3b5e2AFs7f | Abstract
Recent years have witnessed the widespread use of artificial intelligence (AI) algorithms and machine learning (ML) models. Despite their tremendous success, a number of vital problems like ML model brittleness, their fairness, and the lack of interpretability warrant the need for the active developments in e... | [
"{\"IS_META_REVIEW\": false, \"comments\": \"SUMMARY:\\nThe paper proposes and studies a notion of feature attribution in which features are scored for a given instance according to the proportion of minimal explanations for that instance in which they participate. Although an exact computation of this scores can b... |
3b9sqxCW1x | 2,023 | NeurIPS 2023 | true | A Data-Free Approach to Mitigate Catastrophic Forgetting in Federated Class Incremental Learning for Vision Tasks | Deep learning models often suffer from forgetting previously learned information when trained on new data. This problem is exacerbated in federated learning (FL), where the data is distributed and can change independently for each user. Many solutions are proposed to resolve this catastrophic forgetting in a centralize... | [
"federated learning",
"class incremental learning",
"generative models",
"data-free",
"continual learning"
] | https://openreview.net/pdf?id=3b9sqxCW1x | A Data-Free Approach to Mitigate Catastrophic Forgetting in Federated Class Incremental Learning for Vision Tasks
Abstract
Deep learning models often suffer from forgetting previously learned information when trained on new data. This problem is exacerbated in federated learning (FL), where the data is distributed an... | [
"{\"IS_META_REVIEW\": false, \"comments\": \"SUMMARY:\\nIn this paper, authors propose a new approach to perform class-incremental learning in federated setting. Their approach uses a generative model trained at the server side which generates synthetic images to be used as a replacement for data corresponding to o... |
3bdXag2rUd | 2,023 | NeurIPS 2023 | true | SoTTA: Robust Test-Time Adaptation on Noisy Data Streams | Test-time adaptation (TTA) aims to address distributional shifts between training and testing data using only unlabeled test data streams for continual model adaptation. However, most TTA methods assume benign test streams, while test samples could be unexpectedly diverse in the wild. For instance, an unseen object or ... | [
"test-time adaptation",
"domain adaptation",
"deep learning",
"machine learning"
] | https://openreview.net/pdf?id=3bdXag2rUd | SoTTA: Robust Test-Time Adaptation on Noisy Data Streams
Abstract
Test-time adaptation (TTA) aims to address distributional shifts between training and testing data using only unlabeled test data streams for continual model adaptation. However, most TTA methods assume benign test streams, while test samples could be ... | [
"{\"IS_META_REVIEW\": false, \"comments\": \"SUMMARY:\\nThe authors point out that model may suffer from non-interest samples while TTA. Existing TTA methods are not robust to these samples. To address these issues, the authors proposed a methods called SoTTA with two key components, input-wise robustness via high-... |
3fd776zKmo | 2,023 | NeurIPS 2023 | true | Mitigating Over-smoothing in Transformers via Regularized Nonlocal Functionals | Transformers have achieved remarkable success in a wide range of natural language processing and computer vision applications. However, the representation capacity of a deep transformer model is degraded due to the over-smoothing issue in which the token representations become identical when the model's depth grows. In... | [
"transformers",
"self-attention",
"total variation",
"nonlocal functionals",
"over-smoothing"
] | https://openreview.net/pdf?id=3fd776zKmo | Mitigating Over-smoothing in Transformers via Regularized Nonlocal Functionals
Abstract
Transformers have achieved remarkable success in a wide range of natural language processing and computer vision applications. However, the representation capacity of a deep transformer model is degraded due to the over-smoothing ... | [
"{\"IS_META_REVIEW\": false, \"comments\": \"SUMMARY:\\nRecent research works have revealed that the over-smoothing issue, a prevalent challenge in Graph Neural Networks, similarly plagues Transformers. Contrary to expectations, the performance of a Transformer model does not invariably improve with increased depth... |
3gamyee9Yh | 2,023 | NeurIPS 2023 | true | QuantSR: Accurate Low-bit Quantization for Efficient Image Super-Resolution | Low-bit quantization in image super-resolution (SR) has attracted copious attention in recent research due to its ability to reduce parameters and operations significantly. However, many quantized SR models suffer from accuracy degradation compared to their full-precision counterparts, especially at ultra-low bit width... | [
"Super Resolution",
"Model Quantization",
"Deep Learning"
] | https://openreview.net/pdf?id=3gamyee9Yh | QuantSR: Accurate Low-bit Quantization for Efficient Image Super-Resolution
Abstract
Low-bit quantization in image super-resolution (SR) has attracted copious attention in recent research due to its ability to reduce parameters and operations significantly. However, many quantized SR models suffer from accuracy degra... | [
"{\"IS_META_REVIEW\": false, \"comments\": \"SUMMARY:\\nThis manuscript is devoted to pushing the super-resolution (SR) models to ultra-low bit-width (2-4 bits). It proposes two methods and pushes the bit-width to ultra-low 2/4 bits with little accuracy loss. Meanwhile, the proposed methods not only boost the accur... |
3gxiOEf2D6 | 2,023 | NeurIPS 2023 | true | Differentiable and Stable Long-Range Tracking of Multiple Posterior Modes | Particle filters flexibly represent multiple posterior modes nonparametrically, via a collection of weighted samples, but have classically been applied to tracking problems with known dynamics and observation likelihoods. Such generative models may be inaccurate or unavailable for high-dimensional observations like im... | [
"particle",
"filter",
"mixture",
"belief propagation",
"nonparametric",
"deep learning",
"generative",
"discriminative",
"graphical model",
"multiple modes",
"mutli-modal"
] | https://openreview.net/pdf?id=3gxiOEf2D6 | Differentiable and Stable Long-Range Tracking of Multiple Posterior Modes
Abstract
Particle filters flexibly represent multiple posterior modes nonparametrically, via a collection of weighted samples, but have classically been applied to tracking problems with known dynamics and observation likelihoods. Such generati... | [
"{\"IS_META_REVIEW\": false, \"comments\": \"SUMMARY:\\nThe paper proposes the \\\"importance weighted samples gradient (IWSG)\\\" estimator and describes its integration into a \\\"mixture density particle filter (MDPF)\\\" for state space architectures. Similar to regularized particle filters, the MDPF framework ... |
3iSj4l8ZGT | 2,023 | NeurIPS 2023 | true | Learning Interpretable Low-dimensional Representation via Physical Symmetry | We have recently seen great progress in learning interpretable music representations, ranging from basic factors, such as pitch and timbre, to high-level concepts, such as chord and texture. However, most methods rely heavily on music domain knowledge. It remains an open question what general computational principles *... | [
"Physics Symmetry",
"Time series data",
"Self-supervised Learning",
"Representation Augmentation"
] | https://openreview.net/pdf?id=3iSj4l8ZGT | Learning Interpretable Low-dimensional Representation via Physical Symmetry
Abstract
Wehave recently seen great progress in learning interpretable music representations, ranging from basic factors, such as pitch and timbre, to high-level concepts, such as chord and texture. However, most methods rely heavily on music... | [
"{\"IS_META_REVIEW\": false, \"comments\": \"SUMMARY:\\nThis paper presents an approach to interpretable representation learning based on ''physical symmetry''. The core idea is to learn to predict the temporal evolution of a latent variable which is additionally encouraged to be equivariant under some transformati... |
3jAsfo8x8k | 2,023 | NeurIPS 2023 | true | PERFOGRAPH: A Numerical Aware Program Graph Representation for Performance Optimization and Program Analysis | The remarkable growth and significant success of machine learning have expanded its applications into programming languages and program analysis. However, a key challenge in adopting the latest machine learning methods is the representation of programming languages which has a direct impact on the ability of machine le... | [
"program representation",
"graph representation",
"program analysis",
"graph neural networks",
"performance optimization"
] | https://openreview.net/pdf?id=3jAsfo8x8k | PERFOGRAPH: A Numerical Aware Program Graph Representation for Performance Optimization and Program Analysis
Abstract
The remarkable growth and significant success of machine learning have expanded its applications into programming languages and program analysis. However, a key challenge in adopting the latest machin... | [
"{\"IS_META_REVIEW\": false, \"comments\": \"SUMMARY:\\nThe paper proposed a novel graph-based program representation, PerfoGraph, which is based on the current state-of-the-art method PrograML and aims to address its limitation by providing numerical awareness, introducing new tactics for handling local variables,... |
3kitbpEZZO | 2,023 | NeurIPS 2023 | true | Beyond probability partitions: Calibrating neural networks with semantic aware grouping | Research has shown that deep networks tend to be overly optimistic about their predictions, leading to an underestimation of prediction errors. Due to the limited nature of data, existing studies have proposed various methods based on model prediction probabilities to bin the data and evaluate calibration error. We pro... | [
"Uncertainty calibration",
"Deep neural networks"
] | https://openreview.net/pdf?id=3kitbpEZZO | Beyond Probability Partitions: Calibrating Neural Networks with Semantic Aware Grouping
Abstract
Research has shown that deep networks tend to be overly optimistic about their predictions, leading to an underestimation of prediction errors. Due to the limited nature of data, existing studies have proposed various met... | [
"{\"IS_META_REVIEW\": false, \"comments\": \"SUMMARY:\\nThe paper proposes a novel method for dealing with the well-known problem of miscalibration in deep learning models (and machine learning models more generally). The proposed approach is to partition the input space and fit a calibration function to each set. ... |
3n8PNUdvSg | 2,023 | NeurIPS 2023 | true | RECESS Vaccine for Federated Learning: Proactive Defense Against Model Poisoning Attacks | Model poisoning attacks greatly jeopardize the application of federated learning (FL). The effectiveness of existing defenses is susceptible to the latest model poisoning attacks, leading to a decrease in prediction accuracy. Besides, these defenses are intractable to distinguish benign outliers from malicious gradient... | [
"Federated Learning",
"Model Poisoning Attacks",
"Proactive Detection",
"Robust Aggregation",
"Benign Outlier Identification"
] | https://openreview.net/pdf?id=3n8PNUdvSg | RECESS Vaccine for Federated Learning: Proactive Defense Against Model Poisoning Attacks
Abstract
Model poisoning attacks greatly jeopardize the application of federated learning (FL). The effectiveness of existing defenses is susceptible to the latest model poisoning attacks, leading to a decrease in prediction accu... | [
"{\"IS_META_REVIEW\": false, \"comments\": \"SUMMARY:\\nIn this paper, the authors propose RECESS, which is a proactive defense against untargeted model poisoning attacks. Specifically, RECESS proactively detects malicious clients with test gradients and robustly aggregates gradient with a new trust scoring based m... |
3o4jU8fWVj | 2,023 | NeurIPS 2023 | false | EquiformerV2: Improved Equivariant Transformer for Scaling to Higher-Degree Representations | Equivariant Transformers such as Equiformer have demonstrated the efficacy of applying Transformers to the domain of 3D atomistic systems. However, they are still limited to small degrees of equivariant representations due to their computational complexity. In this paper, we investigate whether these architectures can ... | [
"equivariant neural networks",
"graph neural networks",
"computational physics",
"transformer networks"
] | https://openreview.net/pdf?id=3o4jU8fWVj | Abstract
Equivariant Transformers such as Equiformer have demonstrated the efficacy of applying Transformers to the domain of 3D atomistic systems. However, they are still limited to small degrees of equivariant representations due to their computational complexity. In this paper, we investigate whether these architec... | [
"{\"IS_META_REVIEW\": false, \"comments\": \"SUMMARY:\\nThe authors propose EquiformerV2, which is an improvement over the original Equiformer architecture. The main improvement is using a more efficient parameterization of the tensor products used in Equiformer which is computationally expensive for higher order r... |
3ofe0lpwQP | 2,023 | NeurIPS 2023 | true | DisDiff: Unsupervised Disentanglement of Diffusion Probabilistic Models | Targeting to understand the underlying explainable factors behind observations and modeling the conditional generation process on these factors, we connect disentangled representation learning to diffusion probabilistic models (DPMs) to take advantage of the remarkable modeling ability of DPMs. We propose a new task, d... | [
"Diffusion Probabilistic Model",
"Disentangled representation"
] | https://openreview.net/pdf?id=3ofe0lpwQP | DisDiff: Unsupervised Disentanglement of Diffusion Probabilistic Models
Abstract
Targeting to understand the underlying explainable factors behind observations and modeling the conditional generation process on these factors, we connect disentangled representation learning to Diffusion Probabilistic Models (DPMs) to ... | [
"{\"IS_META_REVIEW\": false, \"comments\": \"SUMMARY:\\nImprecisions\\n- “According to the completeness requirement” add reference\\n- line 75-76 PADE-> PDAE\\n- line 270-271. Typo: compare with disco. dissdiff -> compared with disco, disdiff\\n\\nStrengths\\n- Achieving disentangled representations in diffusion mo... |
3pEBW2UPAD | 2,023 | NeurIPS 2023 | true | ReHLine: Regularized Composite ReLU-ReHU Loss Minimization with Linear Computation and Linear Convergence | Empirical risk minimization (ERM) is a crucial framework that offers a general approach to handling a broad range of machine learning tasks. In this paper, we propose a novel algorithm, called ReHLine, for minimizing a set of regularized ERMs with convex piecewise linear-quadratic loss functions and optional linear con... | [
"coordinate descent",
"linear convergence",
"primal-dual methods",
"empirical risk minimization",
"linear constraints",
"quantile regression"
] | https://openreview.net/pdf?id=3pEBW2UPAD | ReHLine: Regularized Composite ReLU-ReHU Loss Minimization with Linear Computation and Linear Convergence
Abstract
Empirical risk minimization (ERM) is a crucial framework that offers a general approach to handling a broad range of machine learning tasks. In this paper, we propose a novel algorithm, called ReHLine, f... | [
"{\"IS_META_REVIEW\": false, \"comments\": \"SUMMARY:\\nThis paper proposes a dual-coordinate descent solver for a class of ERM (Empricial Risk Minimization) problem with general linear inequality constraint, which is of linear convergence rate and efficient coordinate-update cost of O(n) (n is the number of sample... |
3qHlPqzjM1 | 2,023 | NeurIPS 2023 | true | Projection Regret: Reducing Background Bias for Novelty Detection via Diffusion Models | Novelty detection is a fundamental task of machine learning which aims to detect abnormal (*i.e.* out-of-distribution (OOD)) samples. Since diffusion models have recently emerged as the de facto standard generative framework with surprising generation results, novelty detection via diffusion models has also gained much... | [
"Novelty detection",
"out-of-distribution detection",
"consistency models",
"diffusion models",
"score-based generative models"
] | https://openreview.net/pdf?id=3qHlPqzjM1 | Projection Regret: Reducing Background Bias for Novelty Detection via Diffusion Models
Abstract
Novelty detection is a fundamental task of machine learning which aims to detect abnormal ( i.e. out-of-distribution (OOD)) samples. Since diffusion models have recently emerged as the de facto standard generative framewor... | [
"{\"IS_META_REVIEW\": false, \"comments\": \"SUMMARY:\\nRecent methods have mainly utilized the reconstruction property of in-distribution samples to detect OOD by diffusion models. However, they often suffer from detecting OOD samples that share similar background information to the in-distribution data. Based on ... |
3s6aE1LeiR | 2,023 | NeurIPS 2023 | false | From One to Zero: Causal Zero-Shot Neural Architecture Search by Intrinsic One-Shot Interventional Information | ''Zero-shot'' neural architecture search (ZNAS) is key to achieving real-time neural architecture search. ZNAS comes from ''one-shot'' neural architecture search but searches in a weight-agnostic supernet and consequently largely reduce the search cost.
However, the weight parameters are agnostic in the zero-shot NA... | [
"Neural Architecture Search"
] | https://openreview.net/pdf?id=3s6aE1LeiR | Abstract
'Zero-shot' neural architecture search (ZNAS) is key to achieving real-time neural architecture search. ZNAS comes from 'one-shot' neural architecture search but searches in a weight-agnostic supernet and consequently largely reduce the search cost. However, the weight parameters are agnostic in the zero-shot... | [
"{\"IS_META_REVIEW\": false, \"comments\": \"SUMMARY:\\nThis paper proposes a 0-shot NAS method. The key point is that there are latent factors that can influence the architecture search procedure, making the validation accuracy of one-shot NAS unreliable. The method adopts Gaussian intervention to the data and eva... |
3tbTw2ga8K | 2,023 | NeurIPS 2023 | true | The Quantization Model of Neural Scaling | We propose the Quantization Model of neural scaling laws, explaining both the observed power law dropoff of loss with model and data size, and also the sudden emergence of new capabilities with scale. We derive this model from what we call the Quantization Hypothesis,
where network knowledge and skills are "quantized... | [
"scaling laws",
"emergence",
"language models",
"science of deep learning"
] | https://openreview.net/pdf?id=3tbTw2ga8K | The Quantization Model of Neural Scaling
Abstract
We propose the Quantization Model of neural scaling laws, explaining both the observed power law dropoff of loss with model and data size, and also the sudden emergence of new capabilities with scale. We derive this model from what we call the Quantization Hypothesis ... | [
"{\"IS_META_REVIEW\": false, \"comments\": \"SUMMARY:\\nThis paper proposes a possible mechanism that explains both the phenomenon where the cross-entropy loss of large language models (LLMs) decreases as a power law with respect to the training corpus size, and the phenomenon in which certain capabilities of LLMs ... |
3ucmcMzCXD | 2,023 | NeurIPS 2023 | true | Estimating Noise Correlations Across Continuous Conditions With Wishart Processes | The signaling capacity of a neural population depends on the scale and orientation of its covariance across trials. Estimating this "noise" covariance is challenging and is thought to require a large number of stereotyped trials. New approaches are therefore needed to interrogate the structure of neural noise across ri... | [
"Noise Correlations",
"Wishart Process",
"Variational Inference"
] | https://openreview.net/pdf?id=3ucmcMzCXD | Estimating Noise Correlations Across Continuous Conditions With Wishart Processes
Abstract
The signaling capacity of a neural population depends on the scale and orientation of its covariance across trials. Estimating this 'noise' covariance is challenging and is thought to require a large number of stereotyped trial... | [
"{\"IS_META_REVIEW\": false, \"comments\": \"SUMMARY:\\nThe goal of this paper is to compute the covariance of recorded neurons in a given stimulus condition with a low number of samples. Although the conditions are different, some aspects are shared which justifies the fact that the covariance for a specific condi... |
3xRaWBD2YB | 2,023 | NeurIPS 2023 | false | Polynomial Width is Sufficient for Set Representation with High-dimensional Features | Set representation has become ubiquitous in deep learning for modeling the inductive bias of neural networks that are insensitive to the input order. DeepSets is the most widely used neural network architecture for set representation. It involves embedding each set element into a latent space with dimension $L$, follow... | [
"Set Representation; Permutation Invariance; Permutation Equivariance"
] | https://openreview.net/pdf?id=3xRaWBD2YB | Abstract
Set representation has become ubiquitous in deep learning for modeling the inductive bias of neural networks that are insensitive to the input order. DeepSets is the most widely used neural network architecture for set representation. It involves embedding each set element into a latent space with dimension L... | [
"{\"IS_META_REVIEW\": false, \"comments\": \"SUMMARY:\\nSummary: The paper proves that for symmetric neural networks, specifically DeepSets, there exist exact representations for symmetric functions, where the symmetric embedding layer width can be chosen to be polynomial in the set size and input dimension, rather... |
3xSwxlB0fd | 2,023 | NeurIPS 2023 | true | Uncoupled and Convergent Learning in Two-Player Zero-Sum Markov Games with Bandit Feedback | We revisit the problem of learning in two-player zero-sum Markov games, focusing on developing an algorithm that is *uncoupled*, *convergent*, and *rational*, with non-asymptotic convergence rates to Nash equilibrium. We start from the case of stateless matrix game with bandit feedback as a warm-up, showing an $\tilde{... | [
"two-player zero-sum Markov game",
"last-iterate convergence",
"path convergence",
"learning in games"
] | https://openreview.net/pdf?id=3xSwxlB0fd | Uncoupled and Convergent Learning in Two-Player Zero-Sum Markov Games with Bandit Feedback
Abstract
We revisit the problem of learning in two-player zero-sum Markov games, focusing on developing an algorithm that is uncoupled , convergent , and rational , with nonasymptotic convergence rates to Nash equilibrium. We s... | [
"{\"IS_META_REVIEW\": false, \"comments\": \"SUMMARY:\\nThe paper studies the problem of designing uncoupled learning dynamics that provably converge to Nash equilibria in two-player zero-sum Markov games. As a preliminary result, the paper introduces the first dynamics that converge last iterate in self play in ma... |
40L3viVWQN | 2,023 | NeurIPS 2023 | true | The Pick-to-Learn Algorithm: Empowering Compression for Tight Generalization Bounds and Improved Post-training Performance | Generalization bounds are valuable both for theory and applications. On the one hand, they shed light on the mechanisms that underpin the learning processes; on the other, they certify how well a learned model performs against unseen inputs. In this work we build upon a recent breakthrough in compression theory to dev... | [
"Statistical learning theory",
"Compression theory",
"Generalization bounds"
] | https://openreview.net/pdf?id=40L3viVWQN | The Pick-to-Learn Algorithm: Empowering Compression for Tight Generalization Bounds and Improved Post-training Performance
Abstract
Generalization bounds are valuable both for theory and applications. On the one hand, they shed light on the mechanisms that underpin the learning processes; on the other, they certify h... | [
"{\"IS_META_REVIEW\": false, \"comments\": \"SUMMARY:\\nThis paper introduces a meta-learning algorithm which compresses a training set $D$ of size $|D|$ into a smaller training set $T\\\\subset D$ of size $|T|$. The algorithm can be defined for any *base learner*, i.e. any training strategy which outputs a trained... |
42zI5dyNwc | 2,023 | NeurIPS 2023 | false | A database-based rather than a language model-based natural language processing method | Language models applied to NLP tasks take natural language as the direct modeling object. But we believe that natural language is essentially a way of encoding information, therefore, the object of study for natural language should be the information encoded in language, and the organizational and compositional struct... | [
"sentence generation",
"sentence understanding",
"relative spatial relationship",
"Tree-graph hybrid model."
] | https://openreview.net/pdf?id=42zI5dyNwc | null | [
"{\"IS_META_REVIEW\": false, \"comments\": \"SUMMARY:\\nThis paper proposes a new method for natural language processing. Instead of using language corpus, authors suggest to use database. Sentence generation is a linear schematization of a database-based representation. It is indeed an interesting idea.\\n\\nSTREN... |
43ruO2fMjq | 2,023 | NeurIPS 2023 | true | A Unified Framework for U-Net Design and Analysis | U-Nets are a go-to neural architecture across numerous tasks for continuous signals on a square such as images and Partial Differential Equations (PDE), however their design and architecture is understudied. In this paper, we provide a framework for designing and analysing general U-Net architectures. We present theore... | [
"U-Net",
"ResNet",
"Multi-ResNet",
"Generalised U-Net",
"Wavelets",
"Diffusion models",
"Generative modelling",
"PDE Modelling",
"Image Segmentation"
] | https://openreview.net/pdf?id=43ruO2fMjq | A Unified Framework for U-Net Design and Analysis
Abstract
U-Nets are a go-to neural architecture across numerous tasks for continuous signals on a square such as images and Partial Differential Equations (PDE), however their design and architecture is understudied. In this paper, we provide a framework for designing... | [
"{\"IS_META_REVIEW\": false, \"comments\": \"SUMMARY:\\nThis paper proposes a formal definition of U-nets, a crucial building block of modern deep learning pipelines such as diffusion models.\\nThanks to this definition it is possible to both get theoretical results explaining some of the u-nets behaviours, and gen... |
45RBLZBJid | 2,023 | NeurIPS 2023 | true | Accelerated On-Device Forward Neural Network Training with Module-Wise Descending Asynchronism | On-device learning faces memory constraints when optimizing or fine-tuning on edge devices with limited resources. Current techniques for training deep models on edge devices rely heavily on backpropagation. However, its high memory usage calls for a reassessment of its dominance.
In this paper, we propose forward grad... | [
"asynchronous algorithm",
"one-device learning",
"forward gradient descent",
"directional derivative",
"forward algorithms"
] | https://openreview.net/pdf?id=45RBLZBJid | Accelerated On-Device Forward Neural Network Training with Module-Wise Descending Asynchronism
Abstract
On-device learning faces memory constraints when optimizing or fine-tuning on edge devices with limited resources. Current techniques for training deep models on edge devices rely heavily on backpropagation. Howeve... | [
"{\"IS_META_REVIEW\": false, \"comments\": \"SUMMARY:\\nThe paper focuses on enabling training on memory-constrained platforms. Instead of using conventional backpropagation-based methods, the proposed approach is based on forward gradient descent (FGD), which approximates the gradients through only forward passes.... |
46gYakmj4e | 2,023 | NeurIPS 2023 | true | Unsupervised Protein-Ligand Binding Energy Prediction via Neural Euler's Rotation Equation | Protein-ligand binding prediction is a fundamental problem in AI-driven drug discovery. Previous work focused on supervised learning methods for small molecules where binding affinity data is abundant, but it is hard to apply the same strategy to other ligand classes like antibodies where labelled data is limited. In t... | [
"Energy-based Models",
"Denoising Score Matching",
"Equivariant Neural Networks"
] | https://openreview.net/pdf?id=46gYakmj4e | Unsupervised Protein-Ligand Binding Energy Prediction via Neural Euler's Rotation Equation
Abstract
Protein-ligand binding prediction is a fundamental problem in AI-driven drug discovery. Previous work focused on supervised learning methods for small molecules where binding affinity data is abundant, but it is hard t... | [
"{\"IS_META_REVIEW\": false, \"comments\": \"SUMMARY:\\nIn this paper, the authors developed an energy-based model for unsupervised binding affinity prediction. The energy-based model was trained under SE(3) denoising score matching where the rotation score was predicted by Neural Euler’s Rotation Equation. Experim... |
46x3zvYCyQ | 2,023 | NeurIPS 2023 | true | Zeroth-Order Methods for Nondifferentiable, Nonconvex, and Hierarchical Federated Optimization | Federated learning (FL) has emerged as an enabling framework for communication-efficient decentralized training. We study three broadly applicable problem classes in FL: (i) Nondifferentiable nonconvex federated optimization; (ii) Federated bilevel optimization; (iii) Federated minimax problems. Notably, in an implicit... | [
"Federated Learning",
"Nonsmooth Optimization",
"Nonconvex Optimization",
"Bilevel Optimization"
] | https://openreview.net/pdf?id=46x3zvYCyQ | Zeroth-Order Methods for Nondifferentiable, Nonconvex, and Hierarchical Federated Optimization
Abstract
Federated learning (FL) has emerged as an enabling framework for communicationefficient decentralized training. We study three broadly applicable problem classes in FL: (i) Nondifferentiable nonconvex federated opt... | [
"{\"IS_META_REVIEW\": false, \"comments\": \"SUMMARY:\\nThe authors tackle the problem of nondifferentiable nonconvex locally constrained federated learning (Eq. $FL_{nn}$), and the bilevel variant (Eq. $FL_{bl}$). The minimax settings (Eq. $FL_{mm}$) is a special case of the latter. \\n\\nThe authors provide error... |
492Hfmgejy | 2,023 | NeurIPS 2023 | true | Lightweight Vision Transformer with Bidirectional Interaction | Recent advancements in vision backbones have significantly improved their performance by simultaneously modeling images’ local and global contexts. However, the bidirectional interaction between these two contexts has not been well explored and exploited, which is important in the human visual system. This paper propos... | [
"Vision Transformer",
"Lightweight Vision Backbone",
"Convolution Neural Network"
] | https://openreview.net/pdf?id=492Hfmgejy | Lightweight Vision Transformer with Bidirectional Interaction
Abstract
Recent advancements in vision backbones have significantly improved their performance by simultaneously modeling images' local and global contexts. However, the bidirectional interaction between these two contexts has not been well explored and ex... | [
"{\"IS_META_REVIEW\": false, \"comments\": \"SUMMARY:\\nThe paper proposes a new lightweight ViT structure called FAT. They use a fully adaptive self-attention mechanism for vision transformer to model the local and global information as well as the bidirectional interaction between them in context-aware ways. In a... |
497CevPdOg | 2,023 | NeurIPS 2023 | true | Direct Diffusion Bridge using Data Consistency for Inverse Problems | Diffusion model-based inverse problem solvers have shown impressive performance, but are limited in speed, mostly as they require reverse diffusion sampling starting from noise. Several recent works have tried to alleviate this problem by building a diffusion process, directly bridging the clean and the corrupted for s... | [
"Diffusion models",
"Inverse problems",
"Diffusion bridge"
] | https://openreview.net/pdf?id=497CevPdOg | Direct Diffusion Bridge using Data Consistency for Inverse Problems
Abstract
Diffusion model-based inverse problem solvers have shown impressive performance, but are limited in speed, mostly as they require reverse diffusion sampling starting from noise. Several recent works have tried to alleviate this problem by bu... | [
"{\"IS_META_REVIEW\": false, \"comments\": \"SUMMARY:\\nThis paper focuses on the diffusion model-based inversion problem. The paper first analyzes the current works and unifies them with the Direct Diffusion Bridges. Then, the authors point out that the data consistency is ignored by current works and propose CDDB... |
4AQ4Fnemox | 2,023 | NeurIPS 2023 | true | On the Exploitability of Instruction Tuning | Instruction tuning is an effective technique to align large language models (LLMs) with human intent. In this work, we investigate how an adversary can exploit instruction tuning by injecting specific instruction-following examples into the training data that intentionally changes the model's behavior. For example, an ... | [
"Trustworthy machine learning",
"Large language models",
"Supervised fine-tuning",
"instruction tuning"
] | https://openreview.net/pdf?id=4AQ4Fnemox | On the Exploitability of Instruction Tuning
Abstract
Instruction tuning is an effective technique to align large language models (LLMs) with human intents. In this work, we investigate how an adversary can exploit instruction tuning by injecting specific instruction-following examples into the training data that inte... | [
"{\"IS_META_REVIEW\": false, \"comments\": \"SUMMARY:\\nThe authors propose AutoPoison, an automated data poisoning pipeline.\\nThey demonstrate two types of attacks: content injection (e.g, brand names), and over-refusal attacks.\\nThe authors demonstrate how AutoPoison can change a model’s behavior by poisoning o... |
4AmJVaJ78I | 2,023 | NeurIPS 2023 | true | Block-Coordinate Methods and Restarting for Solving Extensive-Form Games | Coordinate descent methods are popular in machine learning and optimization for their simple sparse updates and excellent practical performance.
In the context of large-scale sequential game solving, these same properties would be attractive, but until now no such methods were known, because the strategy spaces do not... | [
"extensive-form games",
"first-order methods",
"coordinate descent"
] | https://openreview.net/pdf?id=4AmJVaJ78I | Block-Coordinate Methods and Restarting for Solving Extensive-Form Games ∗
Abstract
Coordinate descent methods are popular in machine learning and optimization for their simple sparse updates and excellent practical performance. In the context of large-scale sequential game solving, these same properties would be att... | [
"{\"IS_META_REVIEW\": false, \"comments\": \"SUMMARY:\\nThis work proposes a cyclic coordinate descent method to solve the two-player zero-sum extended form game (EFG) and derives the convergence.\\n\\nSTRENGTHS:\\nTo me, solving problems with non-separable constraints by coordinate-descent-type methods is novel an... |
4IWJZjbRFj | 2,023 | NeurIPS 2023 | true | Removing Hidden Confounding in Recommendation: A Unified Multi-Task Learning Approach | In recommender systems, the collected data used for training is always subject to selection bias, which poses a great challenge for unbiased learning. Previous studies proposed various debiasing methods based on observed user and item features, but ignored the effect of hidden confounding. To address this problem, rece... | [
"Debiased recommender system",
"Multi-task learning",
"Causal inference"
] | https://openreview.net/pdf?id=4IWJZjbRFj | Removing Hidden Confounding in Recommendation: A Unified Multi-Task Learning Approach
Abstract
In recommender systems, the collected data used for training is always subject to selection bias, which poses a great challenge for unbiased learning. Previous studies proposed various debiasing methods based on observed us... | [
"{\"IS_META_REVIEW\": false, \"comments\": \"SUMMARY:\\nThe paper addresses that there are hidden confounders in recommendations and existing methods cannot handle them well. \\\\\\nThe paper proposes a unified multi-task learning approach to tackle that problem. \\\\\\nSpecifically, they devise a residual network ... |
4ImZxqmT1K | 2,023 | NeurIPS 2023 | true | Learning to Receive Help: Intervention-Aware Concept Embedding Models | Concept Bottleneck Models (CBMs) tackle the opacity of neural architectures by constructing and explaining their predictions using a set of high-level concepts. A special property of these models is that they permit concept interventions, wherein users can correct mispredicted concepts and thus improve the model's perf... | [
"Explainable Artificial Intelligence",
"Concept Bottleneck Models",
"Concept-based Explainability",
"Interpretability",
"XAI",
"Concept Interventions"
] | https://openreview.net/pdf?id=4ImZxqmT1K | Learning to Receive Help:
Abstract
Concept Bottleneck Models (CBMs) tackle the opacity of neural architectures by constructing and explaining their predictions using a set of high-level concepts. A special property of these models is that they permit concept interventions , wherein users can correct mispredicted conc... | [
"{\"IS_META_REVIEW\": false, \"comments\": \"SUMMARY:\\nThe intervenability of concept bottleneck models has been taken for granted thus far. There have been numerous methods that tried intervening or propose policies to intervene. However, previous methods did not go so far as to optimize models for interventions.... |
4JB42GBxGs | 2,023 | NeurIPS 2023 | true | Neural approximation of Wasserstein distance via a universal architecture for symmetric and factorwise group invariant functions | Learning distance functions between complex objects, such as the Wasserstein distance to compare point sets, is a common goal in machine learning applications. However, functions on such complex objects (e.g., point sets and graphs) are often required to be invariant to a wide variety of group actions e.g. permutation ... | [
"neural networks",
"Wasserstein distance",
"universal approximation",
"optimal transport"
] | https://openreview.net/pdf?id=4JB42GBxGs | Neural approximation of Wasserstein distance via a universal architecture for symmetric and factorwise group invariant functions
Abstract
Learning distance functions between complex objects, such as the Wasserstein distance to compare point sets, is a common goal in machine learning applications. However, functions o... | [
"{\"IS_META_REVIEW\": false, \"comments\": \"SUMMARY:\\nThe authors propose a framework for learning Wasserstien distances and other `SFGI functions'. Two key ingredients in their approach are the characterization as all SFGI functions (a universality theorem), and a sketching mechanism which shows that the size of... |
4JCVw8oMlf | 2,023 | NeurIPS 2023 | true | Effectively Learning Initiation Sets in Hierarchical Reinforcement Learning | An agent learning an option in hierarchical reinforcement learning must solve three problems: identify the option's subgoal (termination condition), learn a policy, and learn where that policy will succeed (initiation set). The termination condition is typically identified first, but the option policy and initiation se... | [
"hierarchical reinforcment learning"
] | https://openreview.net/pdf?id=4JCVw8oMlf | Effectively Learning Initiation Sets in Hierarchical Reinforcement Learning
Abstract
An agent learning an option in hierarchical reinforcement learning must solve three problems: identify the option's subgoal (termination condition), learn a policy, and learn where that policy will succeed (initiation set). The termi... | [
"{\"IS_META_REVIEW\": false, \"comments\": \"SUMMARY:\\nThe paper considers an HRL setting with options (including sub-goal reward functions) in which termination conditions for options are provided, and the goal is to learn the option policies and initiation sets simultaneously. The main focus is on learning of in... |
4KV2xLeqPN | 2,023 | NeurIPS 2023 | true | On the Variance, Admissibility, and Stability of Empirical Risk Minimization | It is well known that Empirical Risk Minimization (ERM) may attain minimax suboptimal rates in terms of the mean squared error (Birgé and Massart, 1993). In this paper, we prove that, under relatively mild assumptions, the suboptimality of ERM must be due to its bias. Namely, the variance error term of ERM (in terms of... | [
"empirical risk minimization",
"bias-variance decomposition",
"admissibility"
] | https://openreview.net/pdf?id=4KV2xLeqPN | On the Variance, Admissibility, and Stability of Empirical Risk Minimization
Abstract
It is well known that Empirical Risk Minimization (ERM) may attain minimax suboptimal rates in terms of the mean squared error (Birgé and Massart, 1993). In this paper, we prove that, under relatively mild assumptions, the suboptima... | [
"{\"IS_META_REVIEW\": false, \"comments\": \"SUMMARY:\\nThis paper proves that variance for ERM enjoys a a minimax rate. The findings indicate that in scenarios where ERM is not optimal, the source of suboptimality lies within the bias component. Furthermore, these insights are extended to encompass an admissibilit... |
4KZhZJSPYU | 2,023 | NeurIPS 2023 | true | When Does Confidence-Based Cascade Deferral Suffice? | Cascades are a classical strategy to enable inference cost to vary adaptively across samples, wherein a sequence of classifiers are invoked in turn. A deferral rule determines whether to invoke the next classifier in the sequence, or to terminate prediction. One simple deferral rule employs the confidence of the curr... | [
"cascades",
"deferral rules",
"adaptive computation",
"model confidence"
] | https://openreview.net/pdf?id=4KZhZJSPYU | When Does Confidence-Based Cascade Deferral Suffice?
Abstract
Cascades are a classical strategy to enable inference cost to vary adaptively across samples, wherein a sequence of classifiers are invoked in turn. A deferral rule determines whether to invoke the next classifier in the sequence, or to terminate predictio... | [
"{\"IS_META_REVIEW\": false, \"comments\": \"SUMMARY:\\nThe paper consists of 2 parts. Part 1 contains theoretical analysis of when confidence-based deferral rules for cascades of 2 or more models succeed or fail, based on a proposed risk function (equation (1) in section 3) presenting a tradeoff between accuracy a... |
4Ks8RPcXd9 | 2,023 | NeurIPS 2023 | true | Direction-oriented Multi-objective Learning: Simple and Provable Stochastic Algorithms | Multi-objective optimization (MOO) has become an influential framework in many machine learning problems with multiple objectives such as learning with multiple criteria and multi-task learning (MTL). In this paper, we propose a new direction-oriented multi-objective formulation by regularizing the common descent direc... | [
"Multi-objective optimization",
"multi-task leaning",
"stochastic algorithms",
"convergence and complexity",
"Pareto stationarity"
] | https://openreview.net/pdf?id=4Ks8RPcXd9 | Direction-oriented Multi-objective Learning: Simple and Provable Stochastic Algorithms
Abstract
Multi-objective optimization (MOO) has become an influential framework in many machine learning problems with multiple objectives such as learning with multiple criteria and multi-task learning (MTL). In this paper, we pro... | [
"{\"IS_META_REVIEW\": false, \"comments\": \"SUMMARY:\\nThis paper proposes a gradient manipulation method named SDMGrad for multi-task learning (MTL). SDMGrad improves the previous MGDA method by using two constraints. The first one is to constrain the common descent direction nearby the one computed with a specif... |
4L2OlXhiTM | 2,023 | NeurIPS 2023 | true | FIRAL: An Active Learning Algorithm for Multinomial Logistic Regression | We investigate theory and algorithms for pool-based active learning for multiclass classification using multinomial logistic regression. Using finite sample analysis, we prove that the Fisher Information Ratio (FIR) lower and upper bounds the excess risk. Based on our theoretical analysis, we propose an active learn... | [
"statistical learning",
"active learning",
"logistic regression",
"regret minimization"
] | https://openreview.net/pdf?id=4L2OlXhiTM | FIRAL: An Active Learning Algorithm for Multinomial Logistic Regression
Abstract
We investigate theory and algorithms for pool-based active learning for multiclass classification using multinomial logistic regression. Using finite sample analysis, we prove that the Fisher Information Ratio (FIR) lower and upper bound... | [
"{\"IS_META_REVIEW\": false, \"comments\": \"SUMMARY:\\nThe authors first prove that the excess risk of multinomial logistic regression with under subgaussian data distribution is lower and upper bounded by terms involving the ratio of the Fisher information of the unlabeled data and the Fisher information of the l... |
4L3RfWnDzL | 2,023 | NeurIPS 2023 | true | Object-centric Learning with Cyclic Walks between Parts and Whole | Learning object-centric representations from complex natural environments enables both humans and machines with reasoning abilities from low-level perceptual features. To capture compositional entities of the scene, we proposed cyclic walks between perceptual features extracted from vision transformers and object entit... | [
"object representation learning",
"slot attention",
"object-centric",
"contrastive random walks"
] | https://openreview.net/pdf?id=4L3RfWnDzL | Object-centric Learning with Cyclic Walks between Parts and Whole
Abstract
Learning object-centric representations from complex natural environments enables both humans and machines with reasoning abilities from low-level perceptual features. To capture compositional entities of the scene, we proposed cyclic walks be... | [
"{\"IS_META_REVIEW\": false, \"comments\": \"SUMMARY:\\nThe paper studies the problem of unsupervised object-centric learning. Previous methods usually leverage reconstruction loss as the supervision signal. The authors propose a novel method that leverages a contrastive cyclic walk loss instead, which was original... |
4L9g1jUDtO | 2,023 | NeurIPS 2023 | true | Generalization in the Face of Adaptivity: A Bayesian Perspective | Repeated use of a data sample via adaptively chosen queries can rapidly lead to overfitting, wherein the empirical evaluation of queries on the sample significantly deviates from their mean with respect to the underlying data distribution. It turns out that simple noise addition algorithms suffice to prevent this issue... | [
"Differential Privacy",
"Adaptive Data Analysis"
] | https://openreview.net/pdf?id=4L9g1jUDtO | Generalization in the Face of Adaptivity: A Bayesian Perspective
Abstract
Repeated use of a data sample via adaptively chosen queries can rapidly lead to overfitting, wherein the empirical evaluation of queries on the sample significantly deviates from their mean with respect to the underlying data distribution. It t... | [
"{\"IS_META_REVIEW\": false, \"comments\": \"SUMMARY:\\nThis paper explores the problem of adaptive data analysis - when a single dataset is repeatedly used for adaptively chosen queries, overfitting can occur rapidly. To reduce this bias, a popular approach is to add noise to the output of each query. Intuitively,... |
4PkBhz18in | 2,023 | NeurIPS 2023 | true | High Precision Causal Model Evaluation with Conditional Randomization | The gold standard for causal model evaluation involves comparing model predictions with true effects estimated from randomized controlled trials (RCT). However, RCTs are not always feasible or ethical to perform. In contrast, conditionally randomized experiments based on inverse probability weighting (IPW) offer a more... | [
"causality",
"causal inference",
"causal model evaluation"
] | https://openreview.net/pdf?id=4PkBhz18in | High Precision Causal Model Evaluation with Conditional Randomization
Abstract
The gold standard for causal model evaluation involves comparing model predictions with true effects estimated from randomized controlled trials (RCT). However, RCTs are not always feasible or ethical to perform. In contrast, conditionally... | [
"{\"IS_META_REVIEW\": false, \"comments\": \"SUMMARY:\\nThe authors formulate and evaluate an approach to solving a non-standard problem: evaluating a causal model M when additional data (not used to construct M) is available from a non-randomized experiment. In particular, the authors focus on comparing IPW estima... |
4R2Y5B12jm | 2,023 | NeurIPS 2023 | true | CRoSS: Diffusion Model Makes Controllable, Robust and Secure Image Steganography | Current image steganography techniques are mainly focused on cover-based methods, which commonly have the risk of leaking secret images and poor robustness against degraded container images. Inspired by recent developments in diffusion models, we discovered that two properties of diffusion models, the ability to achiev... | [
"Diffusion models",
"image steganography",
"Stable Diffusion",
"coverless steganography"
] | https://openreview.net/pdf?id=4R2Y5B12jm | CRoSS: Diffusion Model Makes
Abstract
Current image steganography techniques are mainly focused on cover-based methods, which commonly have the risk of leaking secret images and poor robustness against degraded container images. Inspired by recent developments in diffusion models, we discovered that two properties of... | [
"{\"IS_META_REVIEW\": false, \"comments\": \"SUMMARY:\\nThis paper presents CRoSS, a novel image steganography framework leveraging text-driven diffusion models. It offers improved security, robustness, and controllability compared to cover-based methods.\\n\\nSTRENGTHS:\\nCRoSS is the first work to introduce diffu... |
4RoD1o7yq6 | 2,023 | NeurIPS 2023 | true | Binary Classification with Confidence Difference | Recently, learning with soft labels has been shown to achieve better performance than learning with hard labels in terms of model generalization, calibration, and robustness. However, collecting pointwise labeling confidence for all training examples can be challenging and time-consuming in real-world scenarios. This p... | [
"Weakly supervised learning",
"binary classification",
"unbiased risk estimator"
] | https://openreview.net/pdf?id=4RoD1o7yq6 | Binary Classification with Confidence Difference
Abstract
Recently, learning with soft labels has been shown to achieve better performance than learning with hard labels in terms of model generalization, calibration, and robustness. However, collecting pointwise labeling confidence for all training examples can be ch... | [
"{\"IS_META_REVIEW\": false, \"comments\": \"SUMMARY:\\nThis paper studies binary classification problems. A new data type is introduced, where each observation consists of two input instances, together with the \\\"confidence difference,\\\" defined as the difference of the conditional probabilities (output=1|inpu... |
4SkPTD6XNP | 2,023 | NeurIPS 2023 | true | Cal-DETR: Calibrated Detection Transformer | Albeit revealing impressive predictive performance for several computer vision tasks, deep neural networks (DNNs) are prone to making overconfident predictions. This limits the adoption and wider utilization of DNNs in many safety-critical applications. There have been recent efforts toward calibrating DNNs, however, a... | [
"Model Calibration",
"Object Detection",
"Detection Transformers",
"Uncertainty"
] | https://openreview.net/pdf?id=4SkPTD6XNP | Cal-DETR: Calibrated Detection Transformer
Abstract
Albeit revealing impressive predictive performance for several computer vision tasks, deep neural networks (DNNs) are prone to making overconfident predictions. This limits the adoption and wider utilization of DNNs in many safety-critical applications. There have b... | [
"{\"IS_META_REVIEW\": false, \"comments\": \"SUMMARY:\\nThe paper proposes a method to improve the calibration performance of transformer-based object detectors. In their approach, they first present a way to quantify the uncertainty of each logit using the variance of the outputs of different transformer decoder l... |
4Sn2vUs0zA | 2,023 | NeurIPS 2023 | true | Reference-Based POMDPs | Making good decisions in partially observable and non-deterministic scenarios is a crucial capability for robots. A Partially Observable Markov Decision Process (POMDP) is a general framework for the above problem. Despite advances in POMDP solving, problems with long planning horizons and evolving environments remain ... | [
"POMDP",
"planning under uncertainty",
"long horizon"
] | https://openreview.net/pdf?id=4Sn2vUs0zA | Reference-Based POMDPs
Abstract
Making good decisions in partially observable and non-deterministic scenarios is a crucial capability for robots. A Partially Observable Markov Decision Process (POMDP) is a general framework for the above problem. Despite advances in POMDPsolving, problems with long planning horizons ... | [
"{\"IS_META_REVIEW\": false, \"comments\": \"SUMMARY:\\nThis paper presents a method to solve POMDPs by considering a reference policy. The solution of a reference-based POMDP is presented. Then the existence and uniqueness of the solution are proved. Then the author shows the connections between a reference-based ... |
4SoTUaTK8N | 2,023 | NeurIPS 2023 | true | Reversible and irreversible bracket-based dynamics for deep graph neural networks | Recent works have shown that physics-inspired architectures allow the training of deep graph neural networks (GNNs) without oversmoothing. The role of these physics is unclear, however, with successful examples of both reversible (e.g., Hamiltonian) and irreversible (e.g., diffusion) phenomena producing comparable resu... | [
"graph neural networks",
"structure preserving machine learning",
"neural ordinary differential equations",
"hamiltonian dynamics",
"metriplectic dynamics"
] | https://openreview.net/pdf?id=4SoTUaTK8N | Reversible and irreversible bracket-based dynamics for deep graph neural networks
Abstract
Recent works have shown that physics-inspired architectures allow the training of deep graph neural networks (GNNs) without oversmoothing. The role of these physics is unclear, however, with successful examples of both reversib... | [
"{\"IS_META_REVIEW\": false, \"comments\": \"SUMMARY:\\nThis paper provides a unified framework inspired the bracket-based dynamical system to analysis the oversmoothing problem in GNN. The past work may leverage the opposite physics concept such as the reversible processes, irreversible process and therefore it is... |
4UCktT9XZx | 2,023 | NeurIPS 2023 | true | MuSe-GNN: Learning Unified Gene Representation From Multimodal Biological Graph Data | Discovering genes with similar functions across diverse biomedical contexts poses a significant challenge in gene representation learning due to data heterogeneity. In this study, we resolve this problem by introducing a novel model called Multimodal Similarity Learning Graph Neural Network, which combines Multimodal M... | [
"Multimodal Learning; Representation Learning; Graph Neural Network; Similarity Learning; Contrastive Learning; Computational Biology and Bioinformatics; Single-cell genomics"
] | https://openreview.net/pdf?id=4UCktT9XZx | MuSe-GNN: Learning Unified Gene Representation From Multimodal Biological Graph Data
Abstract
Discovering genes with similar functions across diverse biomedical contexts poses a significant challenge in gene representation learning due to data heterogeneity. In this study, we resolve this problem by introducing a nov... | [
"{\"IS_META_REVIEW\": false, \"comments\": \"SUMMARY:\\nThis paper introduces MuSe-GNN, a model for learning gene embeddings from single-cell sequencing and spatial transcriptomic data that is based on multimodal machine learning and deep graph neural networks. While incorporating regularization with weighted simi... |
4ULTSBBY4U | 2,023 | NeurIPS 2023 | true | Training-free Diffusion Model Adaptation for Variable-Sized Text-to-Image Synthesis | Diffusion models (DMs) have recently gained attention with state-of-the-art performance in text-to-image synthesis. Abiding by the tradition in deep learning, DMs are trained and evaluated on the images with fixed sizes. However, users are demanding for various images with specific sizes and various aspect ratio. This ... | [
"Text-to-Image Synthesis",
"Variable-Sized Image Synthesis",
"Entropy"
] | https://openreview.net/pdf?id=4ULTSBBY4U | Training-free Diffusion Model Adaptation for Variable-Sized Text-to-Image Synthesis
Abstract
Diffusion models (DMs) have recently gained attention with state-of-the-art performance in text-to-image synthesis. Abiding by the tradition in deep learning, DMs are trained and evaluated on the images with fixed sizes. Howe... | [
"{\"IS_META_REVIEW\": false, \"comments\": \"SUMMARY:\\nThis paper proposes both an analysis and a contribution to fix a problem found during the analysis.\\n\\nThey start with the premise that diffusion models should be able to generate arbitrary size images, and training specialized models for each image size is ... |
4VAF3d5jNg | 2,023 | NeurIPS 2023 | true | Adaptive Selective Sampling for Online Prediction with Experts | We consider online prediction of a binary sequence with expert advice. For this setting, we devise label-efficient forecasting algorithms, which use a selective sampling scheme that enables collecting much fewer labels than standard procedures. For the general case without a perfect expert, we prove best-of-both-worlds... | [
"Online learning",
"prediction with experts",
"selective sampling",
"active learning"
] | https://openreview.net/pdf?id=4VAF3d5jNg | Adaptive Selective Sampling for Online Prediction with Experts
Abstract
We consider online prediction of a binary sequence with expert advice. For this setting, we devise label-efficient forecasting algorithms, which use a selective sampling scheme that enables collecting much fewer labels than standard procedures. F... | [
"{\"IS_META_REVIEW\": false, \"comments\": \"SUMMARY:\\nThis paper presents an adaptive label-efficient forecasting technique for online binary prediction with expert advice. The proposed approach implements a label querying probability that is a function of the observed scenario, rather than based on pessimistic c... |
4W9FVg1j6I | 2,023 | NeurIPS 2023 | true | Structured State Space Models for In-Context Reinforcement Learning | Structured state space sequence (S4) models have recently achieved state-of-the-art performance on long-range sequence modeling tasks. These models also have fast inference speeds and parallelisable training, making them potentially useful in many reinforcement learning settings. We propose a modification to a variant... | [
"Reinforcement Learning",
"Meta-Learning",
"State Space Models"
] | https://openreview.net/pdf?id=4W9FVg1j6I | Structured State Space Models for In-Context Reinforcement Learning
Abstract
Structured state space sequence (S4) models have recently achieved state-of-the-art performance on long-range sequence modeling tasks. These models also have fast inference speeds and parallelisable training, making them potentially useful i... | [
"{\"IS_META_REVIEW\": false, \"comments\": \"SUMMARY:\\nThe authors propose a modification to S5 that enables \\\"resetting\\\" the recurrent state, allowing it to function as an RNN replacement in RL. They update the scan operator to utilize the `done` flag and use this to reset the recurrent state. They evaluate ... |
4WPhXYMK6N | 2,023 | NeurIPS 2023 | true | Learning Sample Difficulty from Pre-trained Models for Reliable Prediction | Large-scale pre-trained models have achieved remarkable success in many applications, but how to leverage them to improve the prediction reliability of downstream models is undesirably under-explored. Moreover, modern neural networks have been found to be poorly calibrated and make overconfident predictions regardless ... | [
"uncertainty calibration",
"sample difficulty",
"reliable prediction"
] | https://openreview.net/pdf?id=4WPhXYMK6N | Learning Sample Difficulty from Pre-trained Models for Reliable Prediction
Abstract
Large-scale pre-trained models have achieved remarkable success in many applications, but how to leverage them to improve the prediction reliability of downstream models is undesirably under-explored. Moreover, modern neural networks ... | [
"{\"IS_META_REVIEW\": false, \"comments\": \"SUMMARY:\\nIt is known that state-of-art deep-network based machine learning models are poorly calibrated. It is also somewhat widely known that the poor calibration is because we do not model data uncertainty while training.\\nThis work proposes to use CLIP pretrained m... |
4ZaPpVDjGQ | 2,023 | NeurIPS 2023 | true | Breaking the Communication-Privacy-Accuracy Tradeoff with $f$-Differential Privacy | We consider a federated data analytics problem in which a server coordinates the collaborative data analysis of multiple users with privacy concerns and limited communication capability. The commonly adopted compression schemes introduce information loss into local data while improving communication efficiency, and it ... | [
"Differential privacy",
"federated data analytics",
"discrete valued-mechanism",
"distributed mean estimation"
] | https://openreview.net/pdf?id=4ZaPpVDjGQ | Breaking the Communication-Privacy-Accuracy Tradeoff with f -Differential Privacy
Abstract
We consider a federated data analytics problem in which a server coordinates the collaborative data analysis of multiple users with privacy concerns and limited communication capability. The commonly adopted compression schemes... | [
"{\"IS_META_REVIEW\": false, \"comments\": \"SUMMARY:\\nThis paper studies distributed mean estimation under privacy and communication constraints. This paper focuses on characterizing the recently defined notion of $f$-DP for communication-efficient mechanisms, where $f$-DP can be converted to the standard $(\\\\e... |
4aIpgq1nuI | 2,023 | NeurIPS 2023 | true | What Makes Data Suitable for a Locally Connected Neural Network? A Necessary and Sufficient Condition Based on Quantum Entanglement. | The question of what makes a data distribution suitable for deep learning is a fundamental open problem. Focusing on locally connected neural networks (a prevalent family of architectures that includes convolutional and recurrent neural networks as well as local self-attention models), we address this problem by adopti... | [
"Deep Learning",
"Locally Connected Neural Networks",
"Data Distributions",
"Quantum Entanglement",
"Tensor Networks"
] | https://openreview.net/pdf?id=4aIpgq1nuI | What Makes Data Suitable for a Locally Connected Neural Network? A Necessary and Sufficient Condition Based on Quantum Entanglement
Abstract
The question of what makes a data distribution suitable for deep learning is a fundamental open problem. Focusing on locally connected neural networks (a prevalent family of arc... | [
"{\"IS_META_REVIEW\": false, \"comments\": \"SUMMARY:\\nBy utilizing theoretical tools from quantum physics, the authors propose that a locally connected neural network can accurately predict data if and only if the data distribution exhibits low quantum entanglement under certain feature partitions. Based on this ... |
4anryczeED | 2,023 | NeurIPS 2023 | true | Likelihood Ratio Confidence Sets for Sequential Decision Making | Certifiable, adaptive uncertainty estimates for unknown quantities are an essential ingredient of sequential decision-making algorithms. Standard approaches rely on problem-dependent concentration results and are limited to a specific combination of parameterization, noise family, and estimator. In this paper, we revis... | [
"confidence sets",
"uncertainty quantification",
"bandits",
"active learning",
"testing"
] | https://openreview.net/pdf?id=4anryczeED | Likelihood Ratio Confidence Sets for Sequential Decision Making
Abstract
Certifiable, adaptive uncertainty estimates for unknown quantities are an essential ingredient of sequential decision-making algorithms. Standard approaches rely on problem-dependent concentration results and are limited to a specific combinatio... | [
"{\"IS_META_REVIEW\": false, \"comments\": \"SUMMARY:\\nThis paper proposed to use the likelihood ratio approach to provide an any-time valid confidence sequence, which is suitable for problems with well-specified likelihood. It discusses how to provably choose the best sequence of estimators and sheds light on con... |
4e0NJbkkd8 | 2,023 | NeurIPS 2023 | true | Importance Weighted Actor-Critic for Optimal Conservative Offline Reinforcement Learning | We propose A-Crab (Actor-Critic Regularized by Average Bellman error), a new practical algorithm for offline reinforcement learning (RL) in complex environments with insufficient data coverage. Our algorithm combines the marginalized importance sampling framework with the actor-critic paradigm, where the critic returns... | [
"offline RL",
"actor-critic",
"l_2 single-policy concentrability",
"average bellman error"
] | https://openreview.net/pdf?id=4e0NJbkkd8 | Importance Weighted Actor-Critic for Optimal Conservative Offline Reinforcement Learning
Abstract
We propose A-Crab (Actor-Critic Regularized by Average Bellman error), a new practical algorithm for offline reinforcement learning (RL) in complex environments with insufficient data coverage. Our algorithm combines the... | [
"{\"IS_META_REVIEW\": false, \"comments\": \"SUMMARY:\\nThis paper introduces A-Crab, an offline RL algorithm derived from ATAC, which incorporates a modified loss function for the Q-function. Instead of employing the square Bellman error used in ATAC, A-Crab utilizes the importance-weighted Bellman error. With thi... |
4gLWjSaw4o | 2,023 | NeurIPS 2023 | true | Recovering from Out-of-sample States via Inverse Dynamics in Offline Reinforcement Learning | In this paper we deal with the state distributional shift problem commonly encountered in offline reinforcement learning during test, where the agent tends to take unreliable actions at out-of-sample (unseen) states. Our idea is to encourage the agent to follow the so called state recovery principle when taking actions... | [
"Offline reinforcement learning",
"state distributional shift",
"state recovery",
"inverse dynamics model"
] | https://openreview.net/pdf?id=4gLWjSaw4o | Recovering from Out-of-sample States via Inverse Dynamics in Offline Reinforcement Learning
Abstract
Wedeal with the state distributional shift problem commonly encountered in offline reinforcement learning during test, where the agent tends to take unreliable actions at out-of-sample (unseen) states. Our idea is to ... | [
"{\"IS_META_REVIEW\": false, \"comments\": \"SUMMARY:\\nThis paper aims to tackle a critical challenge in offline reinforcement learning, which involves recovering the state distribution during testing from out-of-sample states. To address this, the authors propose two methods, OSR and OSR-v, which leverage a learn... |
4hYIxI8ds0 | 2,023 | NeurIPS 2023 | true | Make the U in UDA Matter: Invariant Consistency Learning for Unsupervised Domain Adaptation | Domain Adaptation (DA) is always challenged by the spurious correlation between the domain-invariant features (e.g., class identity) and the domain-specific ones (e.g., environment) that does not generalize to the target domain. Unfortunately, even enriched with additional unsupervised target domains, existing Unsuperv... | [
"unsupervised domain adaptation",
"transfer learning"
] | https://openreview.net/pdf?id=4hYIxI8ds0 | Make the U in UDA Matter: Invariant Consistency Learning for Unsupervised Domain Adaptation
Abstract
Domain Adaptation (DA) is always challenged by the spurious correlation between domain-invariant features ( e.g. , class identity) and domain-specific features ( e.g. , environment) that does not generalize to the tar... | [
"{\"IS_META_REVIEW\": false, \"comments\": \"SUMMARY:\\nThis paper presents a novel unsupervised domain adaptation method called invariant CONsistency learning (ICON). ICON is very simple; assuming that labeled source samples and clustered target samples are available, ICON uses BCE losses to make the inner product... |
4hiJ3KPDYe | 2,023 | NeurIPS 2023 | false | Tackling Unconditional Generation for Highly Multimodal Distributions with Hat Diffusion EBM | This work introduces the Hat Diffusion Energy-Based Model (HDEBM), a hybrid of EBMs and diffusion models that can perform high-quality unconditional generation for multimodal image distributions. Our method is motivated by the observation that a partial forward and reverse diffusion defines an MCMC process whose steady... | [
"generative model",
"energy-based model",
"EBM",
"diffusion model",
"hybrid generative model",
"mcmc",
"langevin"
] | https://openreview.net/pdf?id=4hiJ3KPDYe | Z1
ApproximateMCMC
G1(Z1)
Y+G(Z，Z2)
X=
G2(G(Z),Z2)
Energy
H(X;0)
Z1 ~ N(0,1)
Stage-1
Stage-2
Z1 ~ N(0,1)
G1(21)
Z
G2(x,Z2)
G2(x,Z2)
Yo = 0
G
G
update Φ with (10)
K-stepLangevin
p(Y|G(Z,Z2; Φ); 0)
p(Y,Z1,Z2;0)
D
MCMC
X+
r.v.
frozen
H(x;0)
+X
training
H(x; 0)
update θ with (2)
X
memoryba... | [
"{\"IS_META_REVIEW\": false, \"comments\": \"SUMMARY:\\nThis work tries to improve the unconditional generation performance of Energy Based Model (EBM) by combining several techniques. It includes a pertrained diffsion model as a part of the generator and train the energy function and generator through cooperative ... |
4hturzLcKX | 2,023 | NeurIPS 2023 | true | AlpacaFarm: A Simulation Framework for Methods that Learn from Human Feedback | Large language models (LLMs) such as ChatGPT have seen widespread adoption due to their ability to follow user instructions well.
Developing these LLMs involves a complex yet poorly understood workflow requiring training with human feedback. Replicating and understanding this instruction-following process faces three m... | [
"Instruction-Following",
"Reinforcement Learning from Human Feedback",
"Artificial General Intelligence",
"Large Language Models"
] | https://openreview.net/pdf?id=4hturzLcKX | AlpacaFarm: A Simulation Framework for Methods that Learn from Human Feedback
Abstract
Large language models (LLMs) such as ChatGPT have seen widespread adoption due to their ability to follow user instructions well. Developing these LLMs involves a complex yet poorly understood workflow requiring training with human... | [
"{\"IS_META_REVIEW\": false, \"comments\": \"SUMMARY:\\nThis paper presents a simulation framework, AlpacaFarm, for developing LLMs with human feedback.\\nAlpacaFarm adopts LLMs (e.g., GPT-4) to generate feedback (i.e., the ranking of candidate responses given the query), and evaluate the performance by calculating... |
4iMpwAlza1 | 2,023 | NeurIPS 2023 | true | Embroid: Unsupervised Prediction Smoothing Can Improve Few-Shot Classification | Recent work has shown that language models' (LMs) prompt-based learning capabilities make them well suited for automating data labeling in domains where manual annotation is expensive. The challenge is that while writing an initial prompt is cheap, improving a prompt is costly---practitioners often require significant ... | [
"language models",
"prompting",
"embeddings",
"weak supervision"
] | https://openreview.net/pdf?id=4iMpwAlza1 | Embroid: Unsupervised Prediction Smoothing Can Improve Few-Shot Classification
Abstract
Recent work has shown that language models' (LMs) prompt-based learning capabilities make them well suited for automating data labeling in domains where manual annotation is expensive. The challenge is that while writing an initia... | [
"{\"IS_META_REVIEW\": false, \"comments\": \"SUMMARY:\\nThe authors in this paper propose EMBROID aim to improve language models (LM) (such as GPT-3.5) without additional labeled data. Unlike prompt-based models that focus on prompt designs, this paper attempts to modify the predictions of the data point $x$ by con... |
4iTAUsyisM | 2,023 | NeurIPS 2023 | true | Data-Dependent Bounds for Online Portfolio Selection Without Lipschitzness and Smoothness | This work introduces the first small-loss and gradual-variation regret bounds for online portfolio selection, marking the first instances of data-dependent bounds for online convex optimization with non-Lipschitz, non-smooth losses.
The algorithms we propose exhibit sublinear regret rates in the worst cases and achiev... | [
"Online portfolio selection",
"small-loss bound",
"gradual-variation bound",
"second-order bound",
"optimistic FTRL with self-concordant regularizers"
] | https://openreview.net/pdf?id=4iTAUsyisM | Data-Dependent Bounds for Online Portfolio Selection Without Lipschitzness and Smoothness
Abstract
This work introduces the first small-loss and gradual-variation regret bounds for online portfolio selection, marking the first instances of data-dependent bounds for online convex optimization with non-Lipschitz, non-s... | [
"{\"IS_META_REVIEW\": false, \"comments\": \"SUMMARY:\\nThe paper considers the online portfolio selection problem. In this problem, one must allocate funds between d possible investment choices, with the goal of maximizing the total amount. In each round, the \\\"success\\\" of each choice is revealed, in the form... |
4iV26fZPUD | 2,023 | NeurIPS 2023 | true | Train 'n Trade: Foundations of Parameter Markets | Organizations typically train large models individually. This is costly and time-consuming, particularly for large-scale foundation models. Such vertical production is known to be suboptimal. Inspired by this economic insight, we ask whether it is possible to leverage others' expertise by trading the constituent parts ... | [
"Parameter Market",
"Pricing",
"Efficient Model Training"
] | https://openreview.net/pdf?id=4iV26fZPUD | Train 'n Trade: Foundations of Parameter Markets
Abstract
Organizations typically train large models individually. This is costly and time-consuming, particularly for large-scale foundation models. Such vertical production is known to be suboptimal. Inspired by this economic insight, we ask whether it is possible to ... | [
"{\"IS_META_REVIEW\": false, \"comments\": \"SUMMARY:\\nThis paper introduces a novel concept called *parameter markets*, which serves as a platform for exchanging parameters learned by machine learning models. In this framework, agents have the option to engage in parameter trading, to achieve (1) mutual benefits ... |
4jEjq5nhg1 | 2,023 | NeurIPS 2023 | true | Operator Learning with Neural Fields: Tackling PDEs on General Geometries | Machine learning approaches for solving partial differential equations require learning mappings between function spaces. While convolutional or graph neural networks are constrained to discretized functions, neural operators present a promising milestone toward mapping functions directly. Despite impressive results th... | [
"PDEs",
"Physics",
"Operator Learning",
"Deep Learning",
"Spatiotemporal"
] | https://openreview.net/pdf?id=4jEjq5nhg1 | Operator Learning with Neural Fields: Tackling PDEs on General Geometries
Abstract
Machine learning approaches for solving partial differential equations require learning mappings between function spaces. While convolutional or graph neural networks are constrained to discretized functions, neural operators present a... | [
"{\"IS_META_REVIEW\": false, \"comments\": \"SUMMARY:\\nThis paper creatively employs Implicit Neural Representations (INR) for operator learning on irregular domains. This novel method benefits from INR's capability to adaptively handle irregular grid distributions or irregular geometric areas, making learning the... |
4mPiqh4pLb | 2,023 | NeurIPS 2023 | true | Multi-Modal Inverse Constrained Reinforcement Learning from a Mixture of Demonstrations | Inverse Constraint Reinforcement Learning (ICRL) aims to recover the underlying constraints respected by expert agents in a data-driven manner. Existing ICRL algorithms typically assume that the demonstration data is generated by a single type of expert. However, in practice, demonstrations often comprise a mixture of ... | [
"Inverse Constrained Reinforcement Learning",
"Learning from Demonstrations",
"Muti-Modal Learning"
] | https://openreview.net/pdf?id=4mPiqh4pLb | Multi-Modal Inverse Constrained Reinforcement Learning from a Mixture of Demonstrations
Abstract
Inverse Constraint Reinforcement Learning (ICRL) aims to recover the underlying constraints respected by expert agents in a data-driven manner. Existing ICRL algorithms typically assume that the demonstration data is gene... | [
"{\"IS_META_REVIEW\": false, \"comments\": \"SUMMARY:\\nThe paper proposes the algorithm Multi-Modal Inverse Constrained Reinforcement Learning (MMICRL) for imitation learning mixture of expert demonstrations with various constraints. The algorithm includes agent identification, agent-specific constraint inference ... |
4mXYJzoPhf | 2,023 | NeurIPS 2023 | true | Online Pricing for Multi-User Multi-Item Markets | Online pricing has been the focus of extensive research in recent years, particularly in the context of selling an item to sequentially arriving users. However, what if a provider wants to maximize revenue by selling multiple items to multiple users in each round? This presents a complex problem, as the provider must i... | [
"revenue",
"price",
"offer",
"online"
] | https://openreview.net/pdf?id=4mXYJzoPhf | Online Pricing for Multi-User Multi-Item Markets
Abstract
Online pricing has been the focus of extensive research in recent years, particularly in the context of selling an item to sequentially arriving users. However, what if a provider wants to maximize revenue by selling multiple items to multiple users in each ro... | [
"{\"IS_META_REVIEW\": false, \"comments\": \"SUMMARY:\\nThe paper proposes online pricing and allocation strategies for multi-user multi-markets model under three different valuation models.\\n\\nThe main contribution is to extend dynamic pricing strategies from the literature to the case of more than one item and ... |
4mwORQjAim | 2,023 | NeurIPS 2023 | true | Stable Vectorization of Multiparameter Persistent Homology using Signed Barcodes as Measures | Persistent homology (PH) provides topological descriptors for geometric data, such as weighted graphs, which are interpretable, stable to perturbations, and invariant under, e.g., relabeling. Most applications of PH focus on the one-parameter case---where the descriptors summarize the changes in topology of data as it... | [
"topological data analysis",
"multiparameter persistent homology",
"kernel methods",
"optimal transport"
] | https://openreview.net/pdf?id=4mwORQjAim | Stable Vectorization of Multiparameter Persistent Homology using Signed Barcodes as Measures
Abstract
Persistent homology (PH) provides topological descriptors for geometric data, such as weighted graphs, which are interpretable, stable to perturbations, and invariant under, e.g., relabeling. Most applications of PH ... | [
"{\"IS_META_REVIEW\": false, \"comments\": \"SUMMARY:\\nIn this paper, the authors are addressing a very critical need in topological data analysis (TDA), vectorization of multiparameter persistence (MPH). Persistent homology (PH) is the key method in TDA, but in its current form, it allows only a single function t... |
4nSDDokpfK | 2,023 | NeurIPS 2023 | true | Energy-Based Models for Anomaly Detection: A Manifold Diffusion Recovery Approach | We present a new method of training energy-based models (EBMs) for anomaly detection that leverages low-dimensional structures within data. The proposed algorithm, Manifold Projection-Diffusion Recovery (MPDR), first perturbs a data point along a low-dimensional manifold that approximates the training dataset. Then, EB... | [
"Energy-based Models",
"Anomaly Detection",
"Generative Models",
"Out-of-Distribution Detection",
"Recovery Likelihood"
] | https://openreview.net/pdf?id=4nSDDokpfK | Energy-Based Models for Anomaly Detection: A Manifold Diffusion Recovery Approach
Abstract
We present a new method of training energy-based models (EBMs) for anomaly detection that leverages low-dimensional structures within data. The proposed algorithm, Manifold Projection-Diffusion Recovery (MPDR), first perturbs a... | [
"{\"IS_META_REVIEW\": false, \"comments\": \"SUMMARY:\\nThe authors introduce a novel algorithm, Manifold Projection-Diffusion Recovery (MPDR), for training energy-based models (EBMs) that improve the performance of anomaly detection tasks. These tasks are highly relevant in real-world applications like industrial ... |
4qG2RKuZaA | 2,023 | NeurIPS 2023 | true | Censored Sampling of Diffusion Models Using 3 Minutes of Human Feedback | Diffusion models have recently shown remarkable success in high-quality image generation. Sometimes, however, a pre-trained diffusion model exhibits partial misalignment in the sense that the model can generate good images, but it sometimes outputs undesirable images. If so, we simply need to prevent the generation of ... | [
"Generative models",
"Diffusion probabilistic models",
"Controlled generation",
"Human Feedback",
"RLHF"
] | https://openreview.net/pdf?id=4qG2RKuZaA | Censored Sampling of Diffusion Models Using 3 Minutes of Human Feedback
Abstract
Diffusion models have recently shown remarkable success in high-quality image generation. Sometimes, however, a pre-trained diffusion model exhibits partial misalignment in the sense that the model can generate good images, but it someti... | [
"{\"IS_META_REVIEW\": false, \"comments\": \"SUMMARY:\\nThe work aims to solve censored diffusion sampling problem, which prevent diffusion model generating malign / bad images. The core approach is to train a classifier and apply classifier-guided diffusion generation.\\n\\nSTRENGTHS:\\n1. Authors present an inter... |
4sDHLxKb1L | 2,023 | NeurIPS 2023 | true | HiBug: On Human-Interpretable Model Debug | Machine learning models can frequently produce systematic errors on critical subsets (or slices) of data that share common attributes. Discovering and explaining such model bugs is crucial for reliable model deployment. However, existing bug discovery and interpretation methods usually involve heavy human intervention ... | [
"model debugging",
"error slice discovery"
] | https://openreview.net/pdf?id=4sDHLxKb1L | HiBug: On Human-Interpretable Model Debug
Abstract
Machine learning models can frequently produce systematic errors on critical subsets (or slices) of data that share common attributes. Discovering and explaining such model bugs is crucial for reliable model deployment. However, existing bug discovery and interpretat... | [
"{\"IS_META_REVIEW\": false, \"comments\": \"SUMMARY:\\nThis paper proposes a novel model-debugging method for deep learning-based classifiers. Technically, the proposed method first integrates a method for assigning attributes to training data. The method leverages a pre-trained large language model, such as chatG... |
4vGVQVz5KG | 2,023 | NeurIPS 2023 | true | Unsupervised Behavior Extraction via Random Intent Priors | Reward-free data is abundant and contains rich prior knowledge of human behaviors, but it is not well exploited by offline reinforcement learning (RL) algorithms. In this paper, we propose UBER, an unsupervised approach to extract useful behaviors from offline reward-free datasets via diversified rewards. UBER assigns ... | [
"offline RL",
"reward-free",
"behavior extraction"
] | https://openreview.net/pdf?id=4vGVQVz5KG | Unsupervised Behavior Extraction via Random Intent Priors
Abstract
Reward-free data is abundant and contains rich prior knowledge of human behaviors, but it is not well exploited by offline reinforcement learning (RL) algorithms. In this paper, we propose UBER , an unsupervised approach to extract useful behaviors fr... | [
"{\"IS_META_REVIEW\": false, \"comments\": \"SUMMARY:\\nThe authors propose UBER, a method for learning a collection of behavior policies from offline experience data lacking reward labels and ultimately adapting these behaviors in an online setting. UBER generates a collection of randomly-initialized reward models... |
4vpsQdRBlK | 2,023 | NeurIPS 2023 | true | Med-UniC: Unifying Cross-Lingual Medical Vision-Language Pre-Training by Diminishing Bias | The scarcity of data presents a critical obstacle to the efficacy of medical vision-language pre-training (VLP). A potential solution lies in the combination of datasets from various language communities.
Nevertheless, the main challenge stems from the complexity of integrating diverse syntax and semantics, language-sp... | [
"Medical Vision Langauge Pretraining",
"Cross-lingual",
"Language bias"
] | https://openreview.net/pdf?id=4vpsQdRBlK | Med-UniC: Unifying Cross-Lingual Medical Vision-Language Pre-Training by Diminishing Bias
Abstract
The scarcity of data presents a critical obstacle to the efficacy of medical visionlanguage pre-training (VLP). A potential solution lies in the combination of datasets from various language communities. Nevertheless, t... | [
"{\"IS_META_REVIEW\": false, \"comments\": \"SUMMARY:\\nThis paper proposes a vision language pretraining method that focusing on tackling the bias caused by different languages. The Text Alignment Regularization (CTR) is proposed to unify cross-lingual semantic representations of medical reports. The experiments s... |
4xckZu4MPG | 2,023 | NeurIPS 2023 | true | Attention as Implicit Structural Inference | Attention mechanisms play a crucial role in cognitive systems by allowing them to flexibly allocate cognitive resources. Transformers, in particular, have become a dominant architecture in machine learning, with attention as their central innovation. However, the underlying intuition and formalism of attention in Trans... | [
"Attention",
"Structural Inference",
"Variational Inference",
"Predictive Coding",
"Graphical Models"
] | https://openreview.net/pdf?id=4xckZu4MPG | Attention as Implicit Structural Inference
Abstract
Attention mechanisms play a crucial role in cognitive systems by allowing them to flexibly allocate cognitive resources. Transformers, in particular, have become a dominant architecture in machine learning, with attention as their central innovation. However, the un... | [
"{\"IS_META_REVIEW\": false, \"comments\": \"SUMMARY:\\nThe paper shows how attention mechanisms can be interpreted as expectation values over learnable graph connectivity structures given a structural prior; that is from a perspective of (structural) variational inference. The authors first demonstrate this link f... |
4yXnnCK3r9 | 2,023 | NeurIPS 2023 | true | On Proper Learnability between Average- and Worst-case Robustness | Recently, Montasser at al. (2019) showed that finite VC dimension is not sufficient for proper adversarially robust PAC learning. In light of this hardness, there is a growing effort to study what type of relaxations to the adversarially robust PAC learning setup can enable proper learnability. In this work, we initiat... | [
"Adversarial Robustness",
"PAC Learning"
] | https://openreview.net/pdf?id=4yXnnCK3r9 | On Proper Learnability between Average- and Worst-case Robustness
Abstract
Recently, Montasser et al. [2019] showed that finite VC dimension is not sufficient for proper adversarially robust PAC learning. In light of this hardness, there is a growing effort to study what type of relaxations to the adversarially robus... | [
"{\"IS_META_REVIEW\": false, \"comments\": \"SUMMARY:\\nThis paper initiates the study of a new kind of PAC learning: probabilistically robust PAC learning. The authors show that the finiteness of the VC dimension of the function class is not sufficient to obtain a proper learning rule in this new PAC learning set... |
4zWEyYGGfI | 2,023 | NeurIPS 2023 | true | A Unified Detection Framework for Inference-Stage Backdoor Defenses | Backdoor attacks involve inserting poisoned samples during training, resulting in a model containing a hidden backdoor that can trigger specific behaviors without impacting performance on normal samples. These attacks are challenging to detect, as the backdoored model appears normal until activated by the backdoor trig... | [
"Backdoor attacks",
"Backdoor Defense",
"Security for AI"
] | https://openreview.net/pdf?id=4zWEyYGGfI | A Unified Detection Framework for Inference-Stage Backdoor Defenses
Abstract
Backdoor attacks involve inserting poisoned samples during training, resulting in a model containing a hidden backdoor that can trigger specific behaviors without impacting performance on normal samples. These attacks are challenging to dete... | [
"{\"IS_META_REVIEW\": false, \"comments\": \"SUMMARY:\\nThis work formulates the inference-stage backdoor detection problem. The authors then propose a framework to establish provable guarantees w.r.t. the detection FPR, given some validation data on hand. Finally, they derive the optimal detection rule (in the Ney... |
50I7q86igD | 2,023 | NeurIPS 2023 | false | Deep Evidence Regression for Weibull targets | Machine Learning has invariantly found its way into various Credit Risk applications. Due to the intrinsic nature of Credit Risk, quantifying the uncertainty of the predicted risk metrics is essential, and applying uncertainty-aware deep learning models to credit risk settings can be very helpful. In this work, we have... | [
"Deep Learning",
"Probabilistic Methods",
"Computational Finance",
"Credit risk management"
] | https://openreview.net/pdf?id=50I7q86igD | UQ for Credit Risk Management: A deep evidence regression approach
Abstract
11
12
1 Introduction
1.1 Credit Risk Management
Credit risk management is assessing and managing the potential losses that may arise from the 13 failure of borrowers or counterparties to fulfil their financial obligations. In other words,... | [
"{\"IS_META_REVIEW\": false, \"comments\": \"SUMMARY:\\nThis paper aims to explore the application of a scalable UQ-aware deep learning technique, Deep Evidence Regression, and applies it to predict Loss Given Default. It extends the Deep Evidence Regression methodology to learn target variables generated by a Weib... |
50hs53Zb3w | 2,023 | NeurIPS 2023 | true | Recovering Simultaneously Structured Data via Non-Convex Iteratively Reweighted Least Squares | We propose a new algorithm for the problem of recovering data that adheres to multiple, heterogenous low-dimensional structures from linear observations. Focussing on data matrices that are simultaneously row-sparse and low-rank, we propose and analyze an iteratively reweighted least squares (IRLS) algorithm that is ab... | [
"low-rank models",
"sparsity",
"iteratively reweighted least squares",
"non-convex optimization",
"quadratic convergence",
"simultaneously structured data"
] | https://openreview.net/pdf?id=50hs53Zb3w | Recovering Simultaneously Structured Data via Non-Convex Iteratively Reweighted Least Squares
Abstract
We propose a new algorithm for the problem of recovering data that adheres to multiple, heterogeneous low-dimensional structures from linear observations. Focusing on data matrices that are simultaneously row-sparse... | [
"{\"IS_META_REVIEW\": false, \"comments\": \"SUMMARY:\\nThis paper proposes an IRLS method for recovering data with multiple, heterogeneous low-dimensional structures from linear observations. It combines non-convex surrogates for row-sparsity and rank, to identify simultaneously row-sparse and low-rank matrices fr... |
51PLYhMFWz | 2,023 | NeurIPS 2023 | true | Towards a fuller understanding of neurons with Clustered Compositional Explanations | Compositional Explanations is a method for identifying logical formulas of concepts that approximate the neurons' behavior. However, these explanations are linked to the small spectrum of neuron activations (i.e., the highest ones) used to check the alignment, thus lacking completeness. In this paper, we propose a gene... | [
"compositional explanations",
"network dissection",
"explainable artificial intelligence",
"interpretability"
] | https://openreview.net/pdf?id=51PLYhMFWz | Towards a fuller understanding of neurons with Clustered Compositional Explanations
Abstract
Compositional Explanations [30] is a method for identifying logical formulas of concepts that approximate the neurons' behavior. However, these explanations are linked to the small spectrum of neuron activations (i.e., the hi... | [
"{\"IS_META_REVIEW\": false, \"comments\": \"SUMMARY:\\nThis paper is a niche extension of seminal work on network dissection. The authors present a generalization, called Clustered Compositional Explanations, that combines Compositional Explanations with clustering and a novel search heuristic to approximate a bro... |
54hYifmQZU | 2,023 | NeurIPS 2023 | true | Quantifying the Cost of Learning in Queueing Systems | Queueing systems are widely applicable stochastic models with use cases in communication networks, healthcare, service systems, etc.
Although their optimal control has been extensively studied, most existing approaches assume perfect knowledge of the system parameters. Of course, this assumption rarely holds in practi... | [
"bandits",
"learning",
"queueing systems",
"optimal control"
] | https://openreview.net/pdf?id=54hYifmQZU | Quantifying the Cost of Learning in Queueing Systems
Abstract
Queueing systems are widely applicable stochastic models with use cases in communication networks, healthcare, service systems, etc. Although their optimal control has been extensively studied, most existing approaches assume perfect knowledge of the syste... | [
"{\"IS_META_REVIEW\": false, \"comments\": \"SUMMARY:\\nIn this paper, the authors introduce a new regret metric, called Cost of Learning in Queueing (CLQ), to quantify the rate at which an optimal scheduling policy can be learned to minimize the time average queue lengths. The authors derive a lower bound to CLQ a... |
54z8M7NTbJ | 2,023 | NeurIPS 2023 | true | GeoPhy: Differentiable Phylogenetic Inference via Geometric Gradients of Tree Topologies | Phylogenetic inference, grounded in molecular evolution models, is essential for understanding the evolutionary relationships in biological data. Accounting for the uncertainty of phylogenetic tree variables, which include tree topologies and evolutionary distances on branches, is crucial for accurately inferring speci... | [
"phylogenetic inference",
"variational inference",
"control variates",
"hyperbolic space"
] | https://openreview.net/pdf?id=54z8M7NTbJ | GeoPhy: Differentiable Phylogenetic Inference via Geometric Gradients of Tree Topologies
Abstract
Phylogenetic inference, grounded in molecular evolution models, is essential for understanding the evolutionary relationships in biological data. Accounting for the uncertainty of phylogenetic tree variables, which inclu... | [
"{\"IS_META_REVIEW\": false, \"comments\": \"SUMMARY:\\nThis work presents a new robust and scalable method for inferring phylogenetic trees based on (variational) Bayesian inference.\\n\\nSTRENGTHS:\\nOriginality\\n- the originality of this work is in providing a robust and rigorous solution to an important applic... |
559NJBfN20 | 2,023 | NeurIPS 2023 | true | Language Models are Weak Learners | A central notion in practical and theoretical machine learning is that of a *weak learner*, classifiers that achieve better-than-random performance (on any given distribution over data), even by a small margin. Such weak learners form the practical basis for canonical machine learning methods such as boosting. In thi... | [
"language model",
"prompting",
"tabular data",
"summarization",
"boosting",
"adaboost"
] | https://openreview.net/pdf?id=559NJBfN20 | Language models are weak learners
Abstract
A central notion in practical and theoretical machine learning is that of a weak learner , classifiers that achieve better-than-random performance (on any given distribution over data), even by a small margin. Such weak learners form the practical basis for canonical machine... | [
"{\"IS_META_REVIEW\": false, \"comments\": \"SUMMARY:\\nThis paper explored an interesting problem that how to apply and extend LLMs over tabular supervised learning tasks. The paper first described each tabular sample as text, and then resorted to LLM to generate the summary for a set of selected representative sa... |
58HwnnEdtF | 2,023 | NeurIPS 2023 | true | Reward-Directed Conditional Diffusion: Provable Distribution Estimation and Reward Improvement | We explore the methodology and theory of reward-directed generation via conditional diffusion models. Directed generation aims to generate samples with desired properties as measured by a reward function, which has broad applications in generative AI, reinforcement learning, and computational biology. We consider the c... | [
"Theory",
"Diffusion Model",
"Reward Optimization",
"Low-dimensional Data",
"Distribution estimation"
] | https://openreview.net/pdf?id=58HwnnEdtF | Reward-Directed Conditional Diffusion: Provable Distribution Estimation and Reward Improvement
Abstract
We explore the methodology and theory of reward-directed generation via conditional diffusion models. Directed generation aims to generate samples with desired properties as measured by a reward function, which has... | [
"{\"IS_META_REVIEW\": false, \"comments\": \"SUMMARY:\\nThe paper addresses conditional generation with reward-conditioned diffusion models. They propose to learn a reward function from a small subset of labeled data. The paper aims to answer an intriguing research question: \\\"How can we reliably estimate the rew... |
58XMiu8kot | 2,023 | NeurIPS 2023 | true | Bayesian Metric Learning for Uncertainty Quantification in Image Retrieval | We propose a Bayesian encoder for metric learning. Rather than relying on neural amortization as done in prior works, we learn a distribution over the network weights with the Laplace Approximation. We first prove that the contrastive loss is a negative log-likelihood on the spherical space. We propose three methods th... | [
"Laplace approximation",
"metric learning",
"uncertainty quantification",
"weight posterior",
"bayesian"
] | https://openreview.net/pdf?id=58XMiu8kot | Bayesian Metric Learning for Uncertainty Quantification in Image Retrieval
Abstract
We propose a Bayesian encoder for metric learning. Rather than relying on neural amortization as done in prior works, we learn a distribution over the network weights with the Laplace Approximation. We first prove that the contrastive... | [
"{\"IS_META_REVIEW\": false, \"comments\": \"SUMMARY:\\nThis paper presents a Laplace approximation-based probabilistic retrieval approach (aka. Bayesian metric learning for image retrieval). The author provides a probabilistic view of the contrastive loss based on the von-Mises Fisher distribution and corrections ... |
593fc38lhN | 2,023 | NeurIPS 2023 | true | Efficient Meta Neural Heuristic for Multi-Objective Combinatorial Optimization | Recently, neural heuristics based on deep reinforcement learning have exhibited promise in solving multi-objective combinatorial optimization problems (MOCOPs). However, they are still struggling to achieve high learning efficiency and solution quality. To tackle this issue, we propose an efficient meta neural heuristi... | [
"neural heuristic",
"meta learning",
"deep reinforcement learning",
"multi-objective combinatorial optimization"
] | https://openreview.net/pdf?id=593fc38lhN | Efficient Meta Neural Heuristic for Multi-Objective Combinatorial Optimization
Abstract
Recently, neural heuristics based on deep reinforcement learning have exhibited promise in solving multi-objective combinatorial optimization problems (MOCOPs). However, they are still struggling to achieve high learning efficienc... | [
"{\"IS_META_REVIEW\": false, \"comments\": \"SUMMARY:\\nThis article presents an efficient neural heuristic method based on meta-learning, referred to as EMNH (Efficient Meta-learning Neural Heuristic), for solving Multi-Objective Combinatorial Optimization Problems (MOCOPs). The authors employ a shared multi-task ... |
59D5vAGhHQ | 2,023 | NeurIPS 2023 | false | Analyzing and Improving Greedy 2-Coordinate Updates For Equality-Constrained Optimization via Steepest Descent in the 1-Norm | We first consider minimizing a smooth function subject to a summation constraint over its variables.
By exploiting a connection between the greedy 2-coordinate update for this problem and equality-constrained steepest descent in the 1-norm, we give a convergence rate for greedy selection that is faster than random sel... | [
"Coordinate descent",
"SVM",
"LIBSVM",
"Steepest descent",
"convex optimization"
] | https://openreview.net/pdf?id=59D5vAGhHQ | Abstract
We consider minimizing a smooth function subject to a summation constraint over its variables. By exploiting a connection between the greedy 2-coordinate update for this problem and equality-constrained steepest descent in the 1-norm, we give a convergence rate for greedy selection under a proximal PolyakŁ oj... | [
"{\"IS_META_REVIEW\": false, \"comments\": \"SUMMARY:\\nThe paper presents new update rules for block-coordinate descent (BCD) methods to minimize a smooth function subject to one linear equality constraint (precisely, all variables must sum to 1) and possibly a box constraint. A popular method to solve large-scale... |
5AMa9fiyJq | 2,023 | NeurIPS 2023 | true | Common Ground in Cooperative Communication | Cooperative communication plays a fundamental role in theories of human-human interaction--cognition, culture, development, language, etc.--as well as human-robot interaction. The core challenge in cooperative communication is the problem of common ground: having enough shared knowledge and understanding to successfull... | [
"Cooperative Communication",
"Common Ground",
"Bayesian Theory"
] | https://openreview.net/pdf?id=5AMa9fiyJq | Common Ground in Cooperative Communication
Abstract
Cooperative communication plays a fundamental role in theories of human-human interaction-cognition, culture, development, language, etc.-as well as human-robot interaction. The core challenge in cooperative communication is the problem of common ground: having enou... | [
"{\"IS_META_REVIEW\": false, \"comments\": \"SUMMARY:\\nThe authors identify the problem of common ground as the core challenge in cooperative communication, where common ground means having enough shared knowledge and understanding to successfully communicate. They argue that prior models of cooperative communicat... |
5B1ZK60jWn | 2,023 | NeurIPS 2023 | true | A Spectral Theory of Neural Prediction and Alignment | The representations of neural networks are often compared to those of biological systems by performing regression between the neural network responses and those measured from biological systems. Many different state-of-the-art deep neural networks yield similar neural predictions, but it remains unclear how to differen... | [
"computational neuroscience",
"neural manifolds",
"neuro-AI",
"statistical physics",
"representational geometry"
] | https://openreview.net/pdf?id=5B1ZK60jWn | A Spectral Theory of Neural Prediction and Alignment
Abstract
The representations of neural networks are often compared to those of biological systems by performing regression between the neural network responses and those measured from biological systems. Many different state-of-the-art deep neural networks yield si... | [
"{\"IS_META_REVIEW\": false, \"comments\": \"SUMMARY:\\nThis submission uses spectral theory and simulation experiments to try to compare and assess the representations of DNNs vs those of biological neural networks.\\n\\nSTRENGTHS:\\nThe paper comes with a fairly extensive review of the literature.\\nThe paper us... |
5BqDSw8r5j | 2,023 | NeurIPS 2023 | true | Adaptive Normalization for Non-stationary Time Series Forecasting: A Temporal Slice Perspective | Deep learning models have progressively advanced time series forecasting due to their powerful capacity in capturing sequence dependence. Nevertheless, it is still challenging to make accurate predictions due to the existence of non-stationarity in real-world data, denoting the data distribution rapidly changes over ti... | [
"Time series forecasting",
"deep learning",
"normalization"
] | https://openreview.net/pdf?id=5BqDSw8r5j | Adaptive Normalization for Non-stationary Time Series Forecasting: A Temporal Slice Perspective
Abstract
Deep learning models have progressively advanced time series forecasting due to their powerful capacity in capturing sequence dependence. Nevertheless, it is still challenging to make accurate predictions due to t... | [
"{\"IS_META_REVIEW\": false, \"comments\": \"SUMMARY:\\nThe paper proposes a normalization technique that works on sliced time-series with the main goal to remove non-stationary behavior of the inputs (and outputs). The paper computes mean and standard deviation of the slides inputs and then normalizes the inputs b... |
5F04bU79eK | 2,023 | NeurIPS 2023 | true | Provable Guarantees for Neural Networks via Gradient Feature Learning | Neural networks have achieved remarkable empirical performance, while the current theoretical analysis is not adequate for understanding their success, e.g., the Neural Tangent Kernel approach fails to capture their key feature learning ability, while recent analyses on feature learning are typically problem-specific. ... | [
"neural networks",
"gradient descent",
"feature learning",
"provable guarantees",
"theoretical analysis"
] | https://openreview.net/pdf?id=5F04bU79eK | Provable Guarantees for Neural Networks via Gradient Feature Learning
Abstract
Neural networks have achieved remarkable empirical performance, while the current theoretical analysis is not adequate for understanding their success, e.g., the Neural Tangent Kernel approach fails to capture their key feature learning ab... | [
"{\"IS_META_REVIEW\": false, \"comments\": \"SUMMARY:\\nThis paper proposes a general framework for analyzing feature learning in two-layer ReLU neural networks. The idea is to consider the class of two-layer ReLU networks with “gradient features”, i.e. features aligned with the gradients of the loss induced by the... |
5Fgdk3hZpb | 2,023 | NeurIPS 2023 | true | Squeeze, Recover and Relabel: Dataset Condensation at ImageNet Scale From A New Perspective | We present a new dataset condensation framework termed Squeeze, Recover and Relabel (SRe$^2$L) that decouples the bilevel optimization of model and synthetic data during training, to handle varying scales of datasets, model architectures and image resolutions for efficient dataset condensation. The proposed method demo... | [
"Dataset Condensation and Distillation",
"ImageNet Scale"
] | https://openreview.net/pdf?id=5Fgdk3hZpb | Squeeze , Recover and Relabel : Dataset Condensation at ImageNet Scale From A New Perspective
Abstract
We present a new dataset condensation framework termed S queeze ( ), Re cover ( ) and Re labe l ( ) ( SRe 2 L ) that decouples the bilevel optimization of model and synthetic data during training, to handle varying ... | [
"{\"IS_META_REVIEW\": false, \"comments\": \"SUMMARY:\\nThe paper addresses the dataset condensation task and proposes a new framework termed Squeeze, Recover and Relabel. In this three step approach, the authors first train a model from scratch to accommodate most of the crucial information from the original datas... |
5Fr8Nwi5KF | 2,023 | NeurIPS 2023 | true | Diffusion Model for Graph Inverse Problems: Towards Effective Source Localization on Complex Networks | Information diffusion problems, such as the spread of epidemics or rumors, are widespread in society. The inverse problems of graph diffusion, which involve locating the sources and identifying the paths of diffusion based on currently observed diffusion graphs, are crucial to controlling the spread of information. The... | [
"Diffusion Model",
"Graph Inverse Problems",
"Source Localization",
"Information Diffusion"
] | https://openreview.net/pdf?id=5Fr8Nwi5KF | Diffusion Model for Graph Inverse Problems: Towards Effective Source Localization on Complex Networks
Abstract
Information diffusion problems, such as the spread of epidemics or rumors, are widespread in society. The inverse problems of graph diffusion, which involve locating the sources and identifying the paths of ... | [
"{\"IS_META_REVIEW\": false, \"comments\": \"SUMMARY:\\nThis work discusses the challenges associated with tracing the origin and path of information diffusion in complex networks, such as those involved in epidemics or rumors. To address these, the authors propose a probabilistic model, DDMSL (Discrete Diffusion M... |
5GmTI4LNqX | 2,023 | NeurIPS 2023 | true | Strong and Precise Modulation of Human Percepts via Robustified ANNs | The visual object category reports of artificial neural networks (ANNs) are notoriously sensitive to tiny, adversarial image perturbations. Because human category reports (aka human percepts) are thought to be insensitive to those same small-norm perturbations -- and locally stable in general -- this argues that ANNs a... | [
"Vision",
"Object Recognition",
"Human",
"Primate",
"Ventral Stream",
"Adversarial Examples",
"Behavior Modulation",
"Behavioral Alignment"
] | https://openreview.net/pdf?id=5GmTI4LNqX | Strong and Precise Modulation of Human Percepts via Robustified ANNs
Abstract
The visual object category reports of artificial neural networks (ANNs) are notoriously sensitive to tiny, adversarial image perturbations. Because human category reports (aka human percepts) are thought to be insensitive to those same smal... | [
"{\"IS_META_REVIEW\": false, \"comments\": \"SUMMARY:\\nThe paper presents a novel approach to find categorical perceptual changes in humans using artificial neural networks (ANNs). Notably, the paper presents compelling evidence that an adversarially trained resnet50 is better at generating these adversarial atta... |
5Gw9YkJkFF | 2,023 | NeurIPS 2023 | true | PAC Learning Linear Thresholds from Label Proportions | Learning from label proportions (LLP) is a generalization of supervised learning in which the training data is available as sets or bags of feature-vectors (instances) along with the average instance-label of each bag. The goal is to train a good instance classifier. While most previous works on LLP have focused on tra... | [
"PAC learning",
"Learning from label proportions",
"Linear thresholds"
] | https://openreview.net/pdf?id=5Gw9YkJkFF | PAC Learning Linear Thresholds from Label Proportions
Abstract
Learning from label proportions (LLP) is a generalization of supervised learning in which the training data is available as sets or bags of feature-vectors (instances) along with the average instance-label of each bag. The goal is to train a good instance... | [
"{\"IS_META_REVIEW\": false, \"comments\": \"SUMMARY:\\nLearning from label proportions allow training data to aggregate into sets of feature vectors with sum or average of their labels for each set as a label. As supervised learning, the goal is to classify test set of instances and minimize error of the classifie... |
5HahZRA0fy | 2,023 | NeurIPS 2023 | true | Computational Guarantees for Doubly Entropic Wasserstein Barycenters | We study the computation of doubly regularized Wasserstein barycenters, a recently introduced family of entropic barycenters governed by inner and outer regularization strengths. Previous research has demonstrated that various regularization parameter choices unify several notions of entropy-penalized barycenters while... | [
"Wasserstein barycenters",
"entropic penalization",
"optimal transport",
"Sinkhorn's algorithm"
] | https://openreview.net/pdf?id=5HahZRA0fy | Computational Guarantees for Doubly Entropic Wasserstein Barycenters via Damped Sinkhorn Iterations
Abstract
We study the computation of doubly regularized Wasserstein barycenters, a recently introduced family of entropic barycenters governed by inner and outer regularization strengths. Previous research has demonstr... | [
"{\"IS_META_REVIEW\": false, \"comments\": \"SUMMARY:\\nThe paper presents an algorithm (damped Sinkhorn) and theoretical convergence guarantees for computing doubly regularized Wasserstein barycenters. The concept of doubly entropic Wasserstein barycenters extends the single entropic regularized barycenters by int... |
5JcKKRX2iH | 2,023 | NeurIPS 2023 | true | GeoTMI: Predicting Quantum Chemical Property with Easy-to-Obtain Geometry via Positional Denoising | As quantum chemical properties have a dependence on their geometries, graph neural networks (GNNs) using 3D geometric information have achieved high prediction accuracy in many tasks. However, they often require 3D geometries obtained from high-level quantum mechanical calculations, which are practically infeasible, li... | [
"Mutual information",
"Easy-to-obtain geometry",
"Denoising",
"3D Graph neural network",
"OC20"
] | https://openreview.net/pdf?id=5JcKKRX2iH | GeoTMI:
Abstract
As quantum chemical properties have a dependence on their geometries, graph neural networks (GNNs) using 3D geometric information have achieved high prediction accuracy in many tasks. However, they often require 3D geometries obtained from high-level quantum mechanical calculations, which are practic... | [
"{\"IS_META_REVIEW\": false, \"comments\": \"SUMMARY:\\nEdit: Updating score from 5 to 6 based on the discussions.\\n\\nThis work presents a variation of denoising autoencoder type model that uses an easy-to-obtain (corrupted) input geometry to predict properties of molecules. The assumption is that the corrupted g... |
5La4Y8BnQw | 2,023 | NeurIPS 2023 | true | Fast Bellman Updates for Wasserstein Distributionally Robust MDPs | Markov decision processes (MDPs) often suffer from the sensitivity issue under model ambiguity. In recent years, robust MDPs have emerged as an effective framework to overcome this challenge. Distributionally robust MDPs extend the robust MDP framework by incorporating distributional information of the uncertain model ... | [
"Markov decision processes",
"distributionally robust optimization"
] | https://openreview.net/pdf?id=5La4Y8BnQw | Fast Bellman Updates for Wasserstein Distributionally Robust MDPs
Abstract
Markov decision processes (MDPs) often suffer from the sensitivity issue under model ambiguity. In recent years, robust MDPs have emerged as an effective framework to overcome this challenge. Distributionally robust MDPs extend the robust MDP ... | [
"{\"IS_META_REVIEW\": false, \"comments\": \"SUMMARY:\\n* tailored algorithms for solving Wasserstein distributionally robust MDPs and fast implementations with $L_1$, $L_2$, $L_\\\\infty$-based Wasserstein distance.\\n\\nSTRENGTHS:\\n* Fastest algorithms (in terms of dependency on N, S, and A - the number of kerne... |
5MG5C5aS6m | 2,023 | NeurIPS 2023 | true | Global Optimality in Bivariate Gradient-based DAG Learning | Recently, a new class of non-convex optimization problems motivated by the statistical problem of learning an acyclic directed graphical model from data has attracted significant interest. While existing work uses standard first-order optimization schemes to solve this problem, proving the global optimality of such app... | [
"global optimization",
"nonconvex optimization",
"graphical models",
"directed acyclic graphs",
"structure learning"
] | https://openreview.net/pdf?id=5MG5C5aS6m | Global Optimality in Bivariate Gradient-based DAG Learning
Abstract
Recently, a new class of non-convex optimization problems motivated by the statistical problem of learning an acyclic directed graphical model from data has attracted significant interest. While existing work uses standard first-order optimization sc... | [
"{\"IS_META_REVIEW\": false, \"comments\": \"SUMMARY:\\nThe authors give a simple optimization algorithm for DAG-learning-inspired optimization problems that avoids the limitations of known techniques.\\n\\nSTRENGTHS:\\nOriginality:\\nWork is original.\\nI particularly liked the reduction from a combinatorial probl... |
5NxJuc0T1P | 2,023 | NeurIPS 2023 | true | Debias Coarsely, Sample Conditionally: Statistical Downscaling through Optimal Transport and Probabilistic Diffusion Models | We introduce a two-stage probabilistic framework for statistical downscaling using unpaired data. Statistical downscaling seeks a probabilistic map to transform low-resolution data from a biased coarse-grained numerical scheme to high-resolution data that is consistent with a high-fidelity scheme. Our framework tackles... | [
"optimal transport",
"probabilistic diffusion models",
"statistical downscaling"
] | https://openreview.net/pdf?id=5NxJuc0T1P | Debias Coarsely, Sample Conditionally: Statistical Downscaling through Optimal Transport and Probabilistic Diffusion Models
Abstract
We introduce a two-stage probabilistic framework for statistical downscaling using unpaired data . Statistical downscaling seeks a probabilistic map to transform lowresolution data from... | [
"{\"IS_META_REVIEW\": false, \"comments\": \"SUMMARY:\\nThis paper proposed a new two-stage method for statistical downscaling by combining a coarse de-biasing step based on optimal transport and a conditional up-sampling step based on a diffusion model.\\n\\nSTRENGTHS:\\nOverall, the paper is very well written and... |
5R9bZlpZKj | 2,023 | NeurIPS 2023 | true | Smoothed Analysis of Sequential Probability Assignment | We initiate the study of smoothed analysis for the sequential probability assignment problem with contexts. We study information-theoretically optimal minmax rates as well as a framework for algorithmic reduction involving the maximum likelihood estimator oracle. Our approach establishes a general-purpose reduction fro... | [
"Online learning",
"Log loss",
"Information theory",
"Smoothed Analysis",
"Beyond worst case analysis",
"Oracle Efficient Online Learning"
] | https://openreview.net/pdf?id=5R9bZlpZKj | Smoothed Analysis of Sequential Probability Assignment
Abstract
We initiate the study of smoothed analysis for the sequential probability assignment problem with contexts. We study information-theoretically optimal minmax rates as well as a framework for algorithmic reduction involving the maximum likelihood estimato... | [
"{\"IS_META_REVIEW\": false, \"comments\": \"SUMMARY:\\nThis paper focuses on the contextual sequential probability assignments, and specifically examines cases where the contexts $x_{1:T}$ are generated by $\\\\sigma$-smooth adversaries as introduced in [Haghtalab et al. 2021], and where the labels $y_{1:T}$ are a... |
5SIz31OGFV | 2,023 | NeurIPS 2023 | true | Inconsistency, Instability, and Generalization Gap of Deep Neural Network Training | As deep neural networks are highly expressive, it is important to find solutions with small generalization gap (the difference between the performance on the training data and unseen data). Focusing on the stochastic nature of training, we first present a theoretical analysis in which the bound of generalization gap d... | [
"Deep neural network training",
"Generalization gap",
"Empirical study"
] | https://openreview.net/pdf?id=5SIz31OGFV | Inconsistency, Instability, and Generalization Gap of Deep Neural Network Training
Abstract
As deep neural networks are highly expressive, it is important to find solutions with small generalization gap (the difference between the performance on the training data and unseen data). Focusing on the stochastic nature of... | [
"{\"IS_META_REVIEW\": false, \"comments\": \"SUMMARY:\\nThis manuscript propose new notions of inconsistency, instability, and information-theoretifc instability based on the output confidence score to estimate the generalization gap of deep neural networks. Theoretical and empirical results are presented and show ... |
5TTV5IZnLL | 2,023 | NeurIPS 2023 | true | Variational Inference with Gaussian Score Matching | Variational inference (VI) is a method to approximate the computationally intractable posterior distributions that arise in
Bayesian statistics. Typically, VI fits a simple parametric distribution to be close to the target posterior, optimizing an appropriate objective such as the evidence lower bound (ELBO). In th... | [
"Variational Inference",
"score matching",
"KL projection",
"polyak stepsize"
] | https://openreview.net/pdf?id=5TTV5IZnLL | Variational Inference with Gaussian Score Matching
Abstract
Variational inference (VI) is a method to approximate the computationally intractable posterior distributions that arise in Bayesian statistics. Typically, VI fits a simple parametric distribution to be close to the target posterior, optimizing an appropriat... | [
"{\"IS_META_REVIEW\": false, \"comments\": \"SUMMARY:\\nThe submission proposes a method for black-box variational inference based that is based on score matching. The method follows an iterative procedure, where at every iteration the variational approximation is updated by first obtaining a single sample and then... |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.