Armando Solar-Lezama

ABSINT-AI: Agentic Heap Abstractions for Abstract Interpretation

Sparling: End-to-End Spatial Concept Learning via Extremely Sparse Activations

Randomly Sampled Language Reasoning Problems Elucidate Limitations of In-Context Learning

ChartRef: Benchmarking Fine-Grained Visual Element Localization in Charts

VLMaterial: Procedural Material Generation with Large Vision-Language Models

Neurosymbolic World Models for Sequential Decision Making

EnCompass: Enhancing Agent Programming with Search Over Program Execution Paths

LiveCodeBench: Holistic and Contamination Free Evaluation of Large Language Models for Code

CALLME: Call Graph Augmentation with Large Language Models for Javascript

Sparling: Learning Latent Representations with Extremely Sparse Activations

Structured World Models From Low-Level Observations