Direct Unlearning Optimization for Robust and Safe Text-to-Image Models
Comments This paper has been accepted for NeurIPS 2024
期刊&会议
NeurIPS
Conference on Neural Information Processing Systems · 会议 · Machine Learning
Towards Scalable and Stable Parallelization of Nonlinear RNNs
Comments 33 pages, 9 figures, NeurIPS 2024
Cost-aware Bayesian Optimization via the Pandora's Box Gittins Index
Journal ref Advances in Neural Information Processing Systems, 2024
SD-Eval: A Benchmark Dataset for Spoken Dialogue Understanding Beyond Words
Comments Accepted to NeurIPS 2024
BLEnD: A Benchmark for LLMs on Everyday Knowledge in Diverse Cultures and Languages
Comments Accepted to NeurIPS 2024 Datasets & Benchmark Track
Simplified and Generalized Masked Diffusion for Discrete Data
Comments NeurIPS 2024. Code is available at: https://github.com/google-deepmind/md4
The surprising efficiency of temporal difference learning for rare event prediction
Comments Final camera-ready version published at NeurIPS 2024. Correct an assumption statement and typos, and change/add a few sentences from the last version
Provably Efficient Reinforcement Learning with Multinomial Logit Function Approximation
Comments NeurIPS 2024; v3 substantially improves the presentation and further illustrates the role of $κ$ in function approximation
Self-supervised Transformation Learning for Equivariant Representations
Comments 38th Conference on Neural Information Processing Systems (NeurIPS 2024)
A Theory of Optimistically Universal Online Learnability for General Concept Classes
Comments NeurIPS 2024
Score-based 3D molecule generation with neural fields
Comments NeurIPS 2024
Causal vs. Anticausal merging of predictors
Comments Presented at the 38th Conference on Neural Information Processing Systems (NeurIPS 2024)
The Surprising Ineffectiveness of Pre-Trained Visual Representations for Model-Based Reinforcement Learning
Comments Published at the 38th Conference on Neural Information Processing Systems (NeurIPS 2024). Project page: https://schneimo.com/pvr4mbrl/
Debiasing Synthetic Data Generated by Deep Generative Models
Comments Accepted for the 38th Conference on Neural Information Processing Systems (NeurIPS 2024), joint first authors
Taming the Long Tail in Human Mobility Prediction
Comments Accepted by NeurIPS 2024
Supervised Kernel Thinning
Comments Published at NeurIPS 2024
Get Rid of Isolation: A Continuous Multi-task Spatio-Temporal Learning Framework
Comments Accepted by NeurIPS 2024
SCOREQ: Speech Quality Assessment with Contrastive Regression
Comments Accepted NeurIPS 2024
An Accelerated Algorithm for Stochastic Bilevel Optimization under Unbounded Smoothness
Comments Accepted by NeurIPS 2024. The code is available at https://github.com/MingruiLiu-ML-Lab/Accelerated-Bilevel-Optimization-Unbounded-Smoothness
Dual Cone Gradient Descent for Training Physics-Informed Neural Networks
Comments The Thirty-eighth Annual Conference on Neural Information Processing Systems, 2024
A Unified Confidence Sequence for Generalized Linear Models, with Applications to Bandits
Comments 39 pages, 2 figures, 2 tables; Accepted to the 38th Conference on Neural Information Processing Systems (NeurIPS 2024) (ver3: minor revisions, code refactoring; ver2: major revision, including new experiments, reorganization, fixing typos in the proofs of ver1, etc)
Better by Default: Strong Pre-Tuned MLPs and Boosted Trees on Tabular Data
Comments NeurIPS 2024. Changes in v3: mention bug in XGBoost results, mention original name of he+5 method. Code is available at github.com/dholzmueller/pytabkit
Particle Semi-Implicit Variational Inference
Comments NeurIPS 2024 Camera ready
Reciprocal Reward Influence Encourages Cooperation From Self-Interested Agents
Comments NeurIPS 2024
Learning Discrete Concepts in Latent Hierarchical Models
Comments NeurIPS 2024
Understanding Emergent Abilities of Language Models from the Loss Perspective
Comments 23 pages, 8 figures. Accepted in NeurIPS 2024
SelectIT: Selective Instruction Tuning for LLMs via Uncertainty-Aware Self-Reflection
Comments Accepted to NeurIPS 2024
MGF: Mixed Gaussian Flow for Diverse Trajectory Prediction
Comments Accepted by Neurips 2024. Code: https://github.com/mulplue/MGF
Learning Optimal Tax Design in Nonatomic Congestion Games
Comments 23 pages. Accepted by Conference on Neural Information Processing Systems (NeurIPS) 2024
RoME: A Robust Mixed-Effects Bandit Algorithm for Optimizing Mobile Health Interventions
Journal ref 38th Conference on Neural Information Processing Systems (NeurIPS 2024)
