Think, Compress, Answer: Reinforcement Learning for Multimodal Grounding via Reasoning Compression
1 version
Profile built from the corpus for this byline.
Author records are still filling in while the Hub is in beta. If this is your page, you'll be able to claim it soon. Spot a mistake? Tell us.