Timezone: »
The dynamics by which neural networks learn and forget examples throughout training has emerged as an object of interest along several threads of research. In particular, researchers have proposed metrics of example hardness based on these dynamics, including (i) the epoch at which examples are first correctly classified; (ii) the number of times their predictions flip during training; and (iii) whether their prediction flips if they are held out. However, an example might be considered hard for several distinct reasons, such as being a member of a rare subpopulation, being mislabeled, or being fundamentally ambiguous in their class. In this paper, we focus on the second-split forgetting time (SSFT): the epoch (if any) after which an original training example is forgotten as the network is fine-tuned on a randomly held out partition of the data. Across multiple benchmark datasets and modalities, we demonstrate that mislabeled examples are forgotten quickly, and seemingly rare examples are forgotten comparatively slowly. By contrast, metrics only considering the first split learning dynamics struggle to differentiate the two. Additionally, the SSFT tends to be robust to the choice of architecture, optimizer, and random seed. From a practical standpoint, the SSFT (i) can help to identify mislabeled samples, the removal of which improves generalization; and (ii) can provide insights about failure modes.
Author Information
Pratyush Maini (Carnegie Mellon University)
Saurabh Garg (Carnegie Mellon University)
Zachary Lipton (Carnegie Mellon University)
Zico Kolter (Carnegie Mellon University / Bosch Center for AI)
More from the Same Authors
-
2021 : Empirical robustification of pre-trained classifiers »
Mohammad Sadegh Norouzzadeh · Wan-Yi Lin · Leonid Boytsov · Leslie Rice · Huan Zhang · Filipe Condessa · Zico Kolter -
2021 : Certified robustness against adversarial patch attacks via randomized cropping »
Wan-Yi Lin · Fatemeh Sheikholeslami · jinghao shi · Leslie Rice · Zico Kolter -
2021 : Beta-CROWN: Efficient Bound Propagation with Per-neuron Split Constraints for Neural Network Robustness Verification »
Shiqi Wang · Huan Zhang · Kaidi Xu · Xue Lin · Suman Jana · Cho-Jui Hsieh · Zico Kolter -
2021 : Assessing Generalization of SGD via Disagreement Rates »
YiDing Jiang · Vaishnavh Nagarajan · Zico Kolter -
2021 : Do You See What I See? A Comparison of Radiologist Eye Gaze to Computer Vision Saliency Maps for Chest X-ray Classification »
Jesse Kim · Helen Zhou · Zachary Lipton -
2022 : Domain Adaptation under Open Set Label Shift »
Saurabh Garg · Sivaraman Balakrishnan · Zachary Lipton -
2022 : Unsupervised Learning under Latent Label Shift »
Pranav Mani · Manley Roberts · Saurabh Garg · Zachary Lipton -
2022 : Improving adversarial robustness via joint classification and multiple explicit detection classes »
Sina Baharlouei · Fatemeh Sheikholeslami · Meisam Razaviyayn · Zico Kolter -
2022 : Agreement-on-the-Line: Predicting the Performance of Neural Networks under Distribution Shift »
Christina Baek · Yiding Jiang · aditi raghunathan · Zico Kolter -
2022 : Counterfactual Metrics for Auditing Black-Box Recommender Systems for Ethical Concerns »
Nil-Jana Akpinar · Liu Leqi · Dylan Hadfield-Menell · Zachary Lipton -
2022 : RiskyZoo: A Library for Risk-Sensitive Supervised Learning »
William Wong · Audrey Huang · Liu Leqi · Kamyar Azizzadenesheli · Zachary Lipton -
2023 : On the Joint Interaction of Models, Data, and Features »
YiDing Jiang · Christina Baek · Zico Kolter -
2023 : Model-tuning Via Prompts Makes NLP Models Adversarially Robust »
Mrigank Raman · Pratyush Maini · Zico Kolter · Zachary Lipton · Danish Pruthi -
2023 : (Almost) Provable Error Bounds Under Distribution Shift via Disagreement Discrepancy »
Elan Rosenfeld · Saurabh Garg -
2023 : Complementary Benefits of Contrastive Learning and Self-Training Under Distribution Shift »
Saurabh Garg · Amrith Setlur · Zachary Lipton · Sivaraman Balakrishnan · Virginia Smith · Aditi Raghunathan -
2023 : Why is SAM Robust to Label Noise? »
Christina Baek · Zico Kolter · Aditi Raghunathan -
2023 : (Almost) Provable Error Bounds Under Distribution Shift via Disagreement Discrepancy »
Elan Rosenfeld · Saurabh Garg -
2023 : Deep Equilibrium Based Neural Operators for Steady-State PDEs »
Tanya Marwah · Ashwini Pokle · Zico Kolter · Zachary Lipton · Jianfeng Lu · Andrej Risteski -
2023 : How to Cope with Gradual Data Drift? »
Rasool Fakoor · Jonas Mueller · Zachary Lipton · Pratik Chaudhari · Alex Smola -
2023 : TMARS: Improving Visual Representations by Circumventing Text Feature Learning »
Pratyush Maini · Sachin Goyal · Zachary Lipton · Zico Kolter · Aditi Raghunathan -
2023 : Bayesian Neural Networks with Domain Knowledge »
Dylan Sam · Rattana Pukdee · Daniel Jeong · Yewon Byun · Zico Kolter -
2023 : Identifying Inequity in Treatment Allocation »
Yewon Byun · Dylan Sam · Zachary Lipton · Bryan Wilder -
2023 : Language Models are Weak Learners »
Hariharan Manikandan · Yiding Jiang · Zico Kolter -
2023 : A Simple and Effective Pruning Approach for Large Language Models »
Mingjie Sun · Zhuang Liu · Anna Bair · Zico Kolter -
2023 : Conditional Diffusion Replay for Continual Learning in Medical Settings »
Yewon Byun · Saurabh Garg · Sanket Vaibhav Mehta · Praveer Singh · Jayashree Kalpathy-cramer · Bryan Wilder · Zachary Lipton -
2023 : One-Step Diffusion Distillation via Deep Equilibrium Models »
Zhengyang Geng · Ashwini Pokle · Zico Kolter -
2023 : Differentially Private Generation of High Fidelity Samples From Diffusion Models »
Vikash Sehwag · Ashwinee Panda · Ashwini Pokle · Xinyu Tang · Saeed Mahloujifar · Mung Chiang · Zico Kolter · Prateek Mittal -
2023 : Understanding prompt engineering does not require rethinking generalization »
Victor Akinwande · Yiding Jiang · Dylan Sam · Zico Kolter -
2023 : SCIS 2023 Panel, The Future of Generalization: Scale, Safety and Beyond »
Maggie Makar · Samuel Bowman · Zachary Lipton · Adam Gleave -
2023 : Prompt-based Generative Replay: A Text-to-Image Approach for Continual Learning in Medical Settings »
Yewon Byun · Saurabh Garg · Sanket Vaibhav Mehta · Jayashree Kalpathy-Cramer · Praveer Singh · Bryan Wilder · Zachary Lipton -
2023 : (Almost) Provable Error Bounds Under Distribution Shift via Disagreement Discrepancy »
Elan Rosenfeld · Saurabh Garg -
2023 : Zico Kolter »
Zico Kolter -
2023 : Formal Verification for Neural Networks with General Nonlinearities via Branch-and-Bound »
Zhouxing Shi · Qirui Jin · Huan Zhang · Zico Kolter · Suman Jana · Cho-Jui Hsieh -
2023 Workshop: 2nd Workshop on Formal Verification of Machine Learning »
Mark Müller · Brendon G. Anderson · Leslie Rice · Zhouxing Shi · Shubham Ugare · Huan Zhang · Martin Vechev · Zico Kolter · Somayeh Sojoudi · Cho-Jui Hsieh -
2023 : Opening Remarks by Prof. Zico Kolter (CMU) »
Zico Kolter -
2023 Oral: Mimetic Initialization of Self-Attention Layers »
Asher Trockman · Zico Kolter -
2023 Poster: Abstracting Imperfect Information Away from Two-Player Zero-Sum Games »
Samuel Sokota · Ryan D'Orazio · Chun Kai Ling · David Wu · Zico Kolter · Noam Brown -
2023 Poster: Neural Network Approximations of PDEs Beyond Linearity: A Representational Perspective »
Tanya Marwah · Zachary Lipton · Jianfeng Lu · Andrej Risteski -
2023 Poster: Can Neural Network Memorization Be Localized? »
Pratyush Maini · Michael Mozer · Hanie Sedghi · Zachary Lipton · Zico Kolter · Chiyuan Zhang -
2023 Poster: RLSbench: Domain Adaptation Under Relaxed Label Shift »
Saurabh Garg · Nick Erickson · University of California James Sharpnack · Alex Smola · Sivaraman Balakrishnan · Zachary Lipton -
2023 Poster: CHiLS: Zero-Shot Image Classification with Hierarchical Label Sets »
Zachary Novack · Julian McAuley · Zachary Lipton · Saurabh Garg -
2023 Poster: Mimetic Initialization of Self-Attention Layers »
Asher Trockman · Zico Kolter -
2022 Workshop: Principles of Distribution Shift (PODS) »
Elan Rosenfeld · Saurabh Garg · Shibani Santurkar · Jamie Morgenstern · Hossein Mobahi · Zachary Lipton · Andrej Risteski -
2022 : Spotlights »
Pratyush Maini · JIVAT NEET KAUR · Anil Palepu · Polina Kirichenko · Revant Teotia -
2022 Workshop: Workshop on Formal Verification of Machine Learning »
Huan Zhang · Leslie Rice · Kaidi Xu · aditi raghunathan · Wan-Yi Lin · Cho-Jui Hsieh · Clark Barrett · Martin Vechev · Zico Kolter -
2022 Poster: A Branch and Bound Framework for Stronger Adversarial Attacks of ReLU Networks »
Huan Zhang · Shiqi Wang · Kaidi Xu · Yihan Wang · Suman Jana · Cho-Jui Hsieh · Zico Kolter -
2022 Spotlight: A Branch and Bound Framework for Stronger Adversarial Attacks of ReLU Networks »
Huan Zhang · Shiqi Wang · Kaidi Xu · Yihan Wang · Suman Jana · Cho-Jui Hsieh · Zico Kolter -
2022 Poster: Supervised Learning with General Risk Functionals »
Liu Leqi · Audrey Huang · Zachary Lipton · Kamyar Azizzadenesheli -
2022 Poster: Communicating via Markov Decision Processes »
Samuel Sokota · Christian Schroeder · Maximilian Igl · Luisa Zintgraf · Phil Torr · Martin Strohmeier · Zico Kolter · Shimon Whiteson · Jakob Foerster -
2022 Spotlight: Supervised Learning with General Risk Functionals »
Liu Leqi · Audrey Huang · Zachary Lipton · Kamyar Azizzadenesheli -
2022 Spotlight: Communicating via Markov Decision Processes »
Samuel Sokota · Christian Schroeder · Maximilian Igl · Luisa Zintgraf · Phil Torr · Martin Strohmeier · Zico Kolter · Shimon Whiteson · Jakob Foerster -
2021 Workshop: A Blessing in Disguise: The Prospects and Perils of Adversarial Machine Learning »
Hang Su · Yinpeng Dong · Tianyu Pang · Eric Wong · Zico Kolter · Shuo Feng · Bo Li · Henry Liu · Dan Hendrycks · Francesco Croce · Leslie Rice · Tian Tian -
2021 : RL Explainability & Interpretability Panel »
Ofra Amir · Finale Doshi-Velez · Alan Fern · Zachary Lipton · Omer Gottesman · Niranjani Prasad -
2021 Poster: Correcting Exposure Bias for Link Recommendation »
Shantanu Gupta · Hao Wang · Zachary Lipton · Yuyang Wang -
2021 Spotlight: Correcting Exposure Bias for Link Recommendation »
Shantanu Gupta · Hao Wang · Zachary Lipton · Yuyang Wang -
2021 Poster: DORO: Distributional and Outlier Robust Optimization »
Runtian Zhai · Chen Dan · Zico Kolter · Pradeep Ravikumar -
2021 Poster: RATT: Leveraging Unlabeled Data to Guarantee Generalization »
Saurabh Garg · Sivaraman Balakrishnan · Zico Kolter · Zachary Lipton -
2021 Spotlight: DORO: Distributional and Outlier Robust Optimization »
Runtian Zhai · Chen Dan · Zico Kolter · Pradeep Ravikumar -
2021 Oral: RATT: Leveraging Unlabeled Data to Guarantee Generalization »
Saurabh Garg · Sivaraman Balakrishnan · Zico Kolter · Zachary Lipton -
2021 Poster: On Proximal Policy Optimization's Heavy-tailed Gradients »
Saurabh Garg · Joshua Zhanson · Emilio Parisotto · Adarsh Prasad · Zico Kolter · Zachary Lipton · Sivaraman Balakrishnan · Ruslan Salakhutdinov · Pradeep Ravikumar -
2021 Poster: Stabilizing Equilibrium Models by Jacobian Regularization »
Shaojie Bai · Vladlen Koltun · Zico Kolter -
2021 Spotlight: On Proximal Policy Optimization's Heavy-tailed Gradients »
Saurabh Garg · Joshua Zhanson · Emilio Parisotto · Adarsh Prasad · Zico Kolter · Zachary Lipton · Sivaraman Balakrishnan · Ruslan Salakhutdinov · Pradeep Ravikumar -
2021 Spotlight: Stabilizing Equilibrium Models by Jacobian Regularization »
Shaojie Bai · Vladlen Koltun · Zico Kolter -
2020 : Invited Talk: Zico Kolter (Q&A) »
Zico Kolter -
2020 : Invited Talk: Zico Kolter »
Zico Kolter -
2020 : Contributed Talk 3: A Unified View of Label Shift Estimation »
Saurabh Garg -
2020 Poster: Adversarial Robustness Against the Union of Multiple Perturbation Models »
Pratyush Maini · Eric Wong · Zico Kolter -
2020 Poster: Combining Differentiable PDE Solvers and Graph Neural Networks for Fluid Flow Prediction »
Filipe de Avila Belbute-Peres · Thomas Economon · Zico Kolter -
2020 Poster: Certified Robustness to Label-Flipping Attacks via Randomized Smoothing »
Elan Rosenfeld · Ezra Winston · Pradeep Ravikumar · Zico Kolter -
2020 Poster: Uncertainty-Aware Lookahead Factor Models for Quantitative Investing »
Lakshay Chauhan · John Alberg · Zachary Lipton -
2020 Poster: Overfitting in adversarially robust deep learning »
Leslie Rice · Eric Wong · Zico Kolter -
2019 Poster: Domain Adaptation with Asymmetrically-Relaxed Distribution Alignment »
Yifan Wu · Ezra Winston · Divyansh Kaushik · Zachary Lipton -
2019 Poster: What is the Effect of Importance Weighting in Deep Learning? »
Jonathon Byrd · Zachary Lipton -
2019 Oral: Domain Adaptation with Asymmetrically-Relaxed Distribution Alignment »
Yifan Wu · Ezra Winston · Divyansh Kaushik · Zachary Lipton -
2019 Oral: What is the Effect of Importance Weighting in Deep Learning? »
Jonathon Byrd · Zachary Lipton -
2019 Poster: Certified Adversarial Robustness via Randomized Smoothing »
Jeremy Cohen · Elan Rosenfeld · Zico Kolter -
2019 Poster: Wasserstein Adversarial Examples via Projected Sinkhorn Iterations »
Eric Wong · Frank R Schmidt · Zico Kolter -
2019 Oral: Wasserstein Adversarial Examples via Projected Sinkhorn Iterations »
Eric Wong · Frank R Schmidt · Zico Kolter -
2019 Oral: Certified Adversarial Robustness via Randomized Smoothing »
Jeremy Cohen · Elan Rosenfeld · Zico Kolter -
2019 Poster: SATNet: Bridging deep learning and logical reasoning using a differentiable satisfiability solver »
Po-Wei Wang · Priya Donti · Bryan Wilder · Zico Kolter -
2019 Poster: Adversarial camera stickers: A physical camera-based attack on deep learning systems »
Juncheng Li · Frank R Schmidt · Zico Kolter -
2019 Oral: SATNet: Bridging deep learning and logical reasoning using a differentiable satisfiability solver »
Po-Wei Wang · Priya Donti · Bryan Wilder · Zico Kolter -
2019 Oral: Adversarial camera stickers: A physical camera-based attack on deep learning systems »
Juncheng Li · Frank R Schmidt · Zico Kolter -
2018 Poster: Detecting and Correcting for Label Shift with Black Box Predictors »
Zachary Lipton · Yu-Xiang Wang · Alexander Smola -
2018 Poster: Born Again Neural Networks »
Tommaso Furlanello · Zachary Lipton · Michael Tschannen · Laurent Itti · Anima Anandkumar -
2018 Oral: Born Again Neural Networks »
Tommaso Furlanello · Zachary Lipton · Michael Tschannen · Laurent Itti · Anima Anandkumar -
2018 Oral: Detecting and Correcting for Label Shift with Black Box Predictors »
Zachary Lipton · Yu-Xiang Wang · Alexander Smola -
2018 Poster: Provable Defenses against Adversarial Examples via the Convex Outer Adversarial Polytope »
Eric Wong · Zico Kolter -
2018 Oral: Provable Defenses against Adversarial Examples via the Convex Outer Adversarial Polytope »
Eric Wong · Zico Kolter -
2017 Poster: Input Convex Neural Networks »
Brandon Amos · Lei Xu · Zico Kolter -
2017 Poster: OptNet: Differentiable Optimization as a Layer in Neural Networks »
Brandon Amos · Zico Kolter -
2017 Poster: A Semismooth Newton Method for Fast, Generic Convex Programming »
Alnur Ali · Eric Wong · Zico Kolter -
2017 Talk: OptNet: Differentiable Optimization as a Layer in Neural Networks »
Brandon Amos · Zico Kolter -
2017 Talk: Input Convex Neural Networks »
Brandon Amos · Lei Xu · Zico Kolter -
2017 Talk: A Semismooth Newton Method for Fast, Generic Convex Programming »
Alnur Ali · Eric Wong · Zico Kolter