Fine-grained multimodal intelligence requires models to both understand a designated region and localize it from language. Existing methods learn these directions with separate objectives and task-specific annotations. We present Generalized CycleGRPO, the first unified reinforcement learning framework coupling region...
Xin Zhang, Hao-Chen Wang, Bei-Bei Lin et al.· 0 citations
Extending vision-language models (VLMs) to long-form video understanding is bottlenecked by the massive volume of visual tokens that incur quadratic prefill cost in the LLM decoder. Existing compression methods either discard tokens before the question is known (query-agnostic) or passively observe post-hoc attention p...
Dian-Cheng Kang, Ze-Hai Niu, Jiayi Lyu et al.· Proceedings of the 34th ACM...· 0 citations
This work proposes attentional prototype inference (API), a probabilistic latent variable framework for few-shot segmentation, which is formulated as a variational Bayesian inference problem, which is established by amortized inference networks.
Haoliang Sun, Xian-Kai Lu, Hao-Chen Wang et al.· 0 citations