1 paper · 2 filters
Jintao Lin, Bowen Dong, Weikang Shi +4
The capability of Unified Multimodal Models (UMMs) to apply world knowledge across diverse tasks remains a critical, unresolved challenge. Existing benchmarks fall short, offering…