Haolin Yang
PulseAugur coverage of Haolin Yang — every cluster mentioning Haolin Yang across labs, papers, and developer communities, ranked by signal.
-
SoccerMaster: New vision foundation model unifies diverse soccer understanding tasks
Researchers have introduced SoccerMaster, a novel vision foundation model designed for comprehensive soccer understanding. This model unifies various tasks, from athlete detection to event classification, within a singl…
-
Researchers analyze attention heads to understand in-context learning in LLMs
Researchers have developed a new framework called Task Subspace Logit Attribution (TSLA) to analyze how large language models perform in-context learning. This framework identifies specific attention heads responsible f…
-
Researchers develop learned task vectors for improved in-context learning in LLMs
Researchers have developed a new method for training "Learned Task Vectors" (LTVs) that improve the in-context learning capabilities of large language models. Unlike previous methods that extracted task vectors, LTVs ar…
-
StreamAgent anticipates future video events for real-time understanding
Researchers have introduced StreamAgent, a novel system designed for real-time understanding of streaming video. Unlike existing methods that react to events or operate asynchronously, StreamAgent proactively anticipate…