Silvio Savarese

MINT-1T is the largest open-source dataset for training multimodal models.

INDICT enhances code safety and helpfulness using dual critique analysis.

APIGen generates diverse, high-quality datasets for function-calling agents.

Studying how language models respond to fictional questions reveals shared characteristics.

A new framework enhances RL simulations for complex scientific data processing.

A framework to improve teamwork among diverse software engineering agents.

xGen-MM enhances multimodal models for better image and text learning.

A new model improves language understanding and reduces misinformation.

A new benchmark tests AI agents in realistic CRM tasks.

CodeTree transforms coding tasks with a structured process and specialized AI agents.

KALE combines images with rich captions for better understanding.