Reproduce “Prompt-guided Disentangled Representation for Action Recognition” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Tianci Wu, Guangming Zhu, Jiang Lu, Siyuan Wang, Ning Wang, Nuoye Xiong, Zhang Liang
ProDA separates a requested action from its complement by using action prompts to guide graph parsing of objects and relations across video frames. Its prompts condition on the action and node features, while a graph network models interactions over time. On Charades validation, ProDA reports 71.1 mAP versus 67.5 for OR2G+NL; differing pretraining means this is not a controlled comparison.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.