Ayush Jain

Training Active Vision Reasoners with Multi-Turn Reinforcement Learning

Grounded Reinforcement Learning for Visual Reasoning

Unifying 2D and 3D Vision-Language Understanding

LOCATE 3D: Real-World Object Localization via Self-Supervised Learning in 3D

RODIN: Injecting 2D Foundational Features to 3D Vision Language Understanding

From Thousands to Billions: 3D Visual Language Grounding via Render-Supervised Distillation from 2D VLMs