Understanding And Improving Compositional Generalization

Media Summary: Understanding and Improving Compositional Generalization Deciphering the Role of Representation Disentanglement in CLIP Models The team discusses how we should communicate our work on

Understanding And Improving Compositional Generalization - Detailed Analysis & Overview

Understanding and Improving Compositional Generalization Deciphering the Role of Representation Disentanglement in CLIP Models The team discusses how we should communicate our work on Abstract: People learn in fast and flexible ways that elude the best artificial neural networks. Once a person learns how to “dax,” ... Invited talk by Amy Zhang (UC Berkeley and Facebook AI Research) on June 7, 2021 at UCL DARK. Abstract: The benefit of ... Chiyuan Zhang, Google Abstract: Deep learning algorithms are well-known to have a propensity for fitting the training data very ...

This is the first panel discussion of the workshop "The Challenge of Compositionality for Artificial Intelligence", organized by Gary ...

Photo Gallery

Understanding and Improving Compositional Generalization | AI2

Investigating Compositional Generalization in CLIP Models- ECCV 2024

Compositional Generalization Part 1 (Concepts and properties)

Towards a mechanistic understanding of compositionality

Roger Grosse - Studying LLM Generalization through Influence Functions

Compositional Generalization Part 2 (Architecture design and training)

2026/01 - Brainstorming How to Demonstrate Benefits of Compositional Models

Scalable Evaluation and Neural Models for Compositional Generalization [NeurIPS 2025]

Brenden Lake | Compositional generalization in minds and machines

Amy Zhang - Exploring Context for Better Generalization in Reinforcement Learning @ UCL DARK

Quantifying and Understanding Memorization in Deep Neural Networks

Compositional Generalization Part 3 (Inference)

View Detailed Profile

Understanding and Improving Compositional Generalization | AI2

Understanding and Improving Compositional Generalization | AI2

Understanding and Improving Compositional Generalization

Investigating Compositional Generalization in CLIP Models- ECCV 2024

Investigating Compositional Generalization in CLIP Models- ECCV 2024

Deciphering the Role of Representation Disentanglement in CLIP Models |

Compositional Generalization Part 1 (Concepts and properties)

Compositional Generalization Part 1 (Concepts and properties)

Document with more details: https://arxiv.org/abs/2102.04225

Towards a mechanistic understanding of compositionality

Towards a mechanistic understanding of compositionality

Lucas Tian, The Rockefeller University.

Roger Grosse - Studying LLM Generalization through Influence Functions

Roger Grosse - Studying LLM Generalization through Influence Functions

"Studying LLM

Compositional Generalization Part 2 (Architecture design and training)

Compositional Generalization Part 2 (Architecture design and training)

Document with more details: https://arxiv.org/abs/2102.04225

2026/01 - Brainstorming How to Demonstrate Benefits of Compositional Models

2026/01 - Brainstorming How to Demonstrate Benefits of Compositional Models

The team discusses how we should communicate our work on

Scalable Evaluation and Neural Models for Compositional Generalization [NeurIPS 2025]

Scalable Evaluation and Neural Models for Compositional Generalization [NeurIPS 2025]

Compositional generalization

Brenden Lake | Compositional generalization in minds and machines

Brenden Lake | Compositional generalization in minds and machines

Abstract: People learn in fast and flexible ways that elude the best artificial neural networks. Once a person learns how to “dax,” ...

Amy Zhang - Exploring Context for Better Generalization in Reinforcement Learning @ UCL DARK

Amy Zhang - Exploring Context for Better Generalization in Reinforcement Learning @ UCL DARK

Invited talk by Amy Zhang (UC Berkeley and Facebook AI Research) on June 7, 2021 at UCL DARK. Abstract: The benefit of ...

Quantifying and Understanding Memorization in Deep Neural Networks

Quantifying and Understanding Memorization in Deep Neural Networks

Chiyuan Zhang, Google Abstract: Deep learning algorithms are well-known to have a propensity for fitting the training data very ...

Compositional Generalization Part 3 (Inference)

Compositional Generalization Part 3 (Inference)

Document with more details: https://arxiv.org/abs/2102.04225

Why Compositionality Matters for Artificial Intelligence

Why Compositionality Matters for Artificial Intelligence

This is the first panel discussion of the workshop "The Challenge of Compositionality for Artificial Intelligence", organized by Gary ...