Pavel Izmailov

发表

Averaging Weights Leads to Wider Optima and Better Generalization pdf

Andrew Gordon Wilson, Dmitry P. Vetrov, Pavel Izmailov, 2018, UAI.

Loss Surfaces, Mode Connectivity, and Fast Ensembling of DNNs

Andrew Gordon Wilson, Dmitry P. Vetrov, Pavel Izmailov, 2018, NeurIPS.

Invertible Convolutional Networks

Wesley J. Maddox, A. G. Wilson, Wesley Maddox, 2019 .

Fast Uncertainty Estimates and Bayesian Model Averaging of DNNs

A. G. Wilson, D. Vetrov, Wesley Maddox, 2018 .

Improving Stability in Deep Reinforcement Learning with Weight Averaging

A. G. Wilson, D. Vetrov, Pavel Izmailov, 2018 .

Learning Invariances in Neural Networks from Training Data

Andrew Gordon Wilson, Marc Finzi, Pavel Izmailov, 2020, NeurIPS.

Bayesian Model Selection, the Marginal Likelihood, and Generalization

Gregory W. Benton, A. G. Wilson, Pavel Izmailov, 2022, ICML.

On Feature Learning in the Presence of Spurious Correlations

Pavel Izmailov, P. Kirichenko, Nate Gruver, 2022, NeurIPS.

Bayesian Deep Learning and a Probabilistic Perspective of Generalization

Pavel Izmailov, A. Wilson, 2020, NeurIPS.

Evaluating Approximate Inference in Bayesian Deep Learning

Andrew Y. K. Foong, S. Vikram, M. Hoffman, 2021, NeurIPS.

FlexiViT: One Model for All Patch Sizes

Ibrahim M. Alabdulmohsin, M. Tschannen, Alexander Kolesnikov, 2022, 2212.08013.

What Are Bayesian Neural Network Posteriors Really Like?

Andrew Gordon Wilson, Sharad Vikram, Matthew D. Hoffman, 2021, ICML.

A Simple Baseline for Bayesian Uncertainty in Deep Learning

Andrew Gordon Wilson, Dmitry P. Vetrov, Pavel Izmailov, 2019, NeurIPS.

Why Normalizing Flows Fail to Detect Out-of-Distribution Data pdf

Andrew Gordon Wilson, Pavel Izmailov, Polina Kirichenko, 2020, NeurIPS.

Learning Invariances in Neural Networks pdf

Andrew Gordon Wilson, Gregory Benton, Marc Finzi, 2020, NeurIPS.

On Uncertainty, Tempering, and Data Augmentation in Bayesian Classification

Wesley J. Maddox, Sanyam Kapoor, Pavel Izmailov, 2022, NeurIPS.

Generalizing Convolutional Neural Networks for Equivariance to Lie Groups on Arbitrary Continuous Data

Pavel Izmailov, Marc Finzi, S. Stanton, 2020, ICML.

Semi-Supervised Learning with Normalizing Flows pdf

Andrew Gordon Wilson, Marc Finzi, Pavel Izmailov, 2019, ICML.

There Are Many Consistent Explanations of Unlabeled Data: Why You Should Average

Andrew Gordon Wilson, Marc Finzi, Pavel Izmailov, 2018, ICLR.

Subspace Inference for Bayesian Deep Learning

Andrew Gordon Wilson, Dmitry Vetrov, Pavel Izmailov, 2019, UAI.

Unsupervised learning of two-component nematicity from STM data on magic angle bilayer graphene

Pavel Izmailov, S. Nadj-Perge, Eun-Ah Kim, 2022, 2203.04449.

Does Knowledge Distillation Really Work? pdf

Andrew Gordon Wilson, Alexander A. Alemi, Pavel Izmailov, 2021, NeurIPS.

Last Layer Re-Training is Sufficient for Robustness to Spurious Correlations

Pavel Izmailov, P. Kirichenko, A. Wilson, 2022, ArXiv.

FlexiViT: One Model for All Patch Sizes

Ibrahim M. Alabdulmohsin, M. Tschannen, Simon Kornblith, 2022, 2023 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR).

Bayesian Model Selection, the Marginal Likelihood, and Generalization

Gregory W. Benton, Pavel Izmailov, Micah Goldblum, 2022, ICML.

Tensor Train decomposition on TensorFlow (T3F)

Alexander Novikov, Pavel Izmailov, Valentin Khrulkov, 2018, J. Mach. Learn. Res..

Dangers of Bayesian Model Averaging under Covariate Shift pdf

Andrew Gordon Wilson, Patrick K. Nicholson, Pavel Izmailov, 2021, NeurIPS.

Simple and Fast Group Robustness by Automatic Feature Reweighting

Pavel Izmailov, A. Wilson, Shi Qiu, 2023, ICML.

Last Layer Re-Training is Sufficient for Robustness to Spurious Correlations

Pavel Izmailov, P. Kirichenko, A. Wilson, 2022, ICLR.

Improving Consistency-Based Semi-Supervised Learning with Weight Averaging pdf

Andrew Gordon Wilson, Marc Finzi, Pavel Izmailov, 2018, ArXiv.

Scalable Gaussian Processes with Billions of Inducing Inputs via Tensor Train Decomposition

Dmitry Kropotov, Alexander Novikov, Pavel Izmailov, 2017, AISTATS.

Weak-to-Strong Generalization: Eliciting Strong Capabilities With Weak Supervision

Adrien Ecoffet, Pavel Izmailov, J. Leike, 2023, ArXiv.