Search papers, labs, and topics across Lattice.
Affiliation:
3
0
4
Resolving visually indistinguishable human micro-actions requires more than raw foundation model scale: hierarchical coarse-to-fine soft fusion paired with targeted candidate reranking pushes long-tailed micro-action recognition to a benchmark-winning 79.99% F1-mean.
Ignoring the evolving nature of worker performance can lead to suboptimal budget allocation, but this new framework adapts to maximize sensing utility in real-time.
Selecting the right LLM under real-world constraints can lead to significant improvements in service quality and resource efficiency, even in unpredictable environments.
