Recent work has shown that asking language models to generate reasoning ...
Programming is a powerful and ubiquitous problem-solving tool. Developin...
Traditional deep generative models of images and other spatial modalitie...
Recent work has shown the ability to learn generative models for 3D shap...
The information bottleneck principle provides an information-theoretic m...
Understanding the nature of representations learned by supervised machin...
Adversarial examples are typically constructed by perturbing an existing...
We study machine learning formulations of inductive program synthesis; t...
Reinforcement learning is concerned with identifying reward-maximizing
b...
We study machine learning formulations of inductive program synthesis; g...
This paper explores the task of translating natural language queries int...