MAGIC-Video uses a multimodal memory graph and narrative chain inside an agentic loop to enable cross-modal, long-horizon reasoning on ultra-long videos, outperforming prior agentic baselines by 5.9-10.1 points on three benchmarks.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
citation-role summary
background 1
citation-polarity summary
fields
cs.CV 1years
2026 1verdicts
UNVERDICTED 1roles
background 1polarities
background 1representative citing papers
citing papers explorer
-
Bridging Modalities, Spanning Time: Structured Memory for Ultra-Long Agentic Video Reasoning
MAGIC-Video uses a multimodal memory graph and narrative chain inside an agentic loop to enable cross-modal, long-horizon reasoning on ultra-long videos, outperforming prior agentic baselines by 5.9-10.1 points on three benchmarks.