Track manipulation, locomotion and embodied systems.
Search papers, labs, and topics across Lattice.
Samsung Research
4
0
6
Robot-specific fine-tuning may not be necessary for complex manipulation: structured in-context memory and execution-receipt grammar allow frozen frontier VLMs to match specialized hybrid VLA policies within two points.
Closing the supervision gap in GUI agents boosts success rates from the low-30% range to over 50% through innovative skill-guided learning.
Current mobile GUI agents are surprisingly inept at everyday smartphone tasks, achieving only 62% success on a new benchmark of real-world Android apps.