Neural Networks And Spurious Correlations
Aditi Raghunathan (UC Berkeley)
https://simons.berkeley.edu/talks/neural-networks-and-spurious-correlations
Algorithmic Aspects of Causal Inference
Neural networks trained on large datasets have powered several notable successes across various domains. However, such models often fail in the presence of distribution shifts because they can pick up on spurious correlations present in the dataset. In this talk, I will discuss how to train neural networks that are robust to known spurious correlations. We will see that this leads to counter-intuitive observations about the effect of model size and training data. Finally, I will discuss some approaches to mitigating the effect of spurious correlations in the absence of complete information about the spurious correlations.
Видео Neural Networks And Spurious Correlations канала Simons Institute
https://simons.berkeley.edu/talks/neural-networks-and-spurious-correlations
Algorithmic Aspects of Causal Inference
Neural networks trained on large datasets have powered several notable successes across various domains. However, such models often fail in the presence of distribution shifts because they can pick up on spurious correlations present in the dataset. In this talk, I will discuss how to train neural networks that are robust to known spurious correlations. We will see that this leads to counter-intuitive observations about the effect of model size and training data. Finally, I will discuss some approaches to mitigating the effect of spurious correlations in the absence of complete information about the spurious correlations.
Видео Neural Networks And Spurious Correlations канала Simons Institute
Показать
Комментарии отсутствуют
Информация о видео
Другие видео канала
Berkeley in the 80s, Episode 5: Richard KarpAdversarial Examples in Deep LearningProbabilistic Programming for Augmented IntelligenceQuantum Advantage Without StructureRecent Advances in Quantum Simulation with Applications to Chemistry and Field TheoryRemarks on the Discrete CubeFast Reinforcement Learning With Generalized Policy Updates"The Problem with Qubits"Stability and Learning in Repeated GamesMachine Learning-based Design of Proteins and Small MoleculesExact Asymptotics and Universality for Gradient Flows and Empirical Risk MinimizersRecent Developments in Supervised Learning With NoiseRates of Normal Approximation for Typical Weighted SumsPlanning and Markov Decision Processes Part 1 (reupload)Reductionism in Reinforcement LearningSubmodular OptimizationTractable Learning in Structured Probability SpacesUsing Molecular Networks to Seed Predictive Hierarchies of Cell SystemsPanel | Quantum ColloquiumTractable Probabilistic CircuitsAlgorithmic Fairness From The Lens Of Causality And Information Theory