Search papers, labs, and topics across Lattice.
1
0
2
Existing MLLMs struggle with procedural reasoning, but a new self-skill-exploration agent achieves state-of-the-art performance by discovering effective strategies for everyday tasks.
