Paper recorded by Signals 4 on 2026-08-31 in cs.CV. Abstract reproduced from arXiv; link to the original below.
Published 2026-08-31 on arXiv · recorded by Signals 4 on 2026-09-01
Category: cs.CV · 计算机视觉 · first seen 2026-09-01
Existing long-video agents acquire evidence through one uniform behavior, ignoring whether the required evidence is concentrated, requires broad occurrence coverage, or must discriminate competing hypotheses---which can cause failure before substantive reasoning begins. Prescribing a fine-grained solution procedure for every question is not a satisfactory remedy, as it restricts autonomous explora