# IMPACT-CYCLE: A Contract-Based Multi-Agent System for Claim-Level Supervisory Correction of Long-Video Semantic Memory

Source: [IMPACT-CYCLE: A Contract-Based Multi-Agent System for Claim-Level Supervisory Correction of Long-Video Semantic Memory](<https://www.sota2.com/research/paper/69e9a7aa55a3c32af89f713d>)

Collection index: [llms.txt](<https://www.sota2.com/research/llms.txt>)

Data updated: 2026-04-24T05:00:04.605Z

Authors: Weitong Kong, Di Wen, Kunyu Peng, David Schneider, Zeyun Zhong, Alexander Jaus, Zdravko Marinov, Jiale Wei, Ruiping Liu, Junwei Zheng, Yufan Chen, Lei Qi, Rainer Stiefelhagen

Published: 2026-04-22T03:03:33.000Z

## Abstract

Correcting errors in long-video understanding is disproportionately costly: existing multimodal pipelines produce opaque, end-to-end outputs that expose no intermediate state for inspection, forcing annotators to revisit raw video and reconstruct temporal logic from scratch. The core bottleneck is not generation quality alone, but the absence of a supervisory interface through which human effort can be proportional to the scope of each error. We present IMPACT-CYCLE, a supervisory multi-agent system that reformulates long-video understanding as iterative claim-level maintenance of a shared semantic memory -- a structured, versioned state encoding typed claims, a claim dependency graph, and a provenance log. Role-specialized agents operating under explicit authority contracts decompose verification into local object-relation correctness, cross-temporal consistency, and global semantic coherence, with corrections confined to structurally dependent claims. When automated evidence is insufficient, the system escalates to human arbitration as the supervisory authority with final override rights; dependency-closure re-verification then ensures correction cost remains proportional to error scope. Experiments on VidOR show substantially improved downstream reasoning (VQA: 0.71 to 0.79) and a 4.8x reduction in human arbitration cost, with workload significantly lower than manual annotation. Code will be released at https://github.com/MKong17/IMPACT\_CYCLE.

## Recorded benchmark results

A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage.

- [Video Question Answering on VidOR](<https://www.sota2.com/research/sota/video-question-answering-on-vidor.md>): Video Question Answering; VidOR; VQA Accuracy: 79.

- [Video Scene Graph Generation on VidOR](<https://www.sota2.com/research/sota/video-scene-graph-generation-on-vidor.md>): Video Scene Graph Generation; VidOR; Entity Accuracy: 93.1.

## Original sources and implementation

[Original paper](<https://arxiv.org/abs/2604.20136>)

[PDF](<https://arxiv.org/pdf/2604.20136.pdf>)
