Baishakhi Ray

Enhancing code generation models to handle minor input changes effectively.

Examining the challenges and future roles in automatic programming with LLMs.

Enhancing LLMs' ability to refine their code through self-debugging techniques.

A fresh method for training code models focusing on semantics and execution behavior.

A new approach to assess reasoning strategies with a focus on computational costs.

New platform and dataset aim to enhance bug-fixing models for the Linux kernel.

A new system enhances Zebra puzzle solving accuracy using AI agents and language models.

Using large language models to detect software weaknesses.