pith. machine review for the scientific record. sign in

Video-o3: Native interleaved clue seeking for long video multi-hop reasoning

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 1

years

2026 1

verdicts

UNVERDICTED 1

roles

background 1

polarities

background 1

representative citing papers

VISD: Enhancing Video Reasoning via Structured Self-Distillation

cs.CV · 2026-05-07 · unverdicted · novelty 5.0 · 3 refs

VISD adds structured privileged feedback from a judge model and a direction-magnitude decoupling trick to let VideoLLMs learn token-level credit assignment while keeping RL stable, yielding higher accuracy and roughly 2x faster convergence on video reasoning benchmarks.

citing papers explorer

Showing 1 of 1 citing paper.

  • VISD: Enhancing Video Reasoning via Structured Self-Distillation cs.CV · 2026-05-07 · unverdicted · none · ref 48 · 3 links

    VISD adds structured privileged feedback from a judge model and a direction-magnitude decoupling trick to let VideoLLMs learn token-level credit assignment while keeping RL stable, yielding higher accuracy and roughly 2x faster convergence on video reasoning benchmarks.