A new online video LLM framework decouples reasoning control from memory via ATDM for evidence-timed transparent responses and HPSI for progressive semantic integration, improving benchmarks to 71.6% on StreamingBench.
Clip X→[supports/contradicts/provides evidence for] [attribute] because [exact caption text]
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CV 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Progressive Online Video Understanding with Evidence-Aligned Timing and Transparent Decisions
A new online video LLM framework decouples reasoning control from memory via ATDM for evidence-timed transparent responses and HPSI for progressive semantic integration, improving benchmarks to 71.6% on StreamingBench.