Search papers, labs, and topics across Lattice.
State Key Laboratory of General Artificial Intelligence, BIGAI
4
0
5
2
Outperforming previous methods, UniLM-Nav achieves zero-shot last-mile navigation by effectively integrating multimodal reasoning and task context.
A single structural edit can drastically impair LLM performance in molecular tasks, highlighting the fragility of their generalization capabilities.
GUI agents struggle in dynamic environments because they only see static screenshots, but DynamicUI's video-based approach with frame selection and action-conditioned refinement leaps ahead.
Existing GUI agents can parrot actions, but AutoGUI-v2 reveals they still lack a deep understanding of GUI functionality and struggle to predict the outcomes of even simple interactions.
