Search papers, labs, and topics across Lattice.
Queen Mary University of London
4
0
5
Tailoring evaluation taxonomies for each vision-language model reveals a 32% improvement in performance and uncovers unique model blind spots that global assessments miss.
Achieving competitive video super-resolution quality with just 11.25% of the usual trainable parameters, LiteVSR redefines efficiency in adapting frozen diffusion models.
User interaction can break your LLM's confidence calibration, but this new method can fix it.
By explicitly modeling event-level causal relationships in videos via scene graphs, GraphThinker significantly reduces hallucinations in video reasoning tasks.
