Yanbin Wei, Chun Kang, Siwei Li, Haoxuan Che, Yang Chen, Hua Liu, Jian Liu, Zhuang Liu, Can Ouyang, Fei Xing, Lei Sha, Rui Liu, Yu Zhang, James Kwok
View original ↗Build a library that visualizes and processes hypergraph data using standard LVLM architectures. This would provide the community with a baseline to test how well models handle complex, non-linear relationships compared to standard graph benchmarks.
Suggested repo: hypervision
"See the hidden connections: First-ever framework for hypergraph-aware LVLM evaluation."
Estimated effort: 40h