EviBridge: Encoding Predictive Visual Evidence for Proactive Interaction
Tool-handover interaction calls for preparation before a response is warranted. The model waits until interaction cues appear, then prepares while assessing whether the accumulating evidence supports a response or calls for cancellation. We introduce EviBridge, a framework that integrates predictive visual representations into a vision-language interaction model for joint state and tool prediction. A trainable evidence interface maps features from a frozen V-JEPA 2 encoder into the language decoder alongside native visual inputs, preserving their temporal context and using only observations available at each decision time. The framework has two instantiations. EviBridge-E distinguishes recent from historical evidence and encodes the timing and spatial location of each observation. EviBridge-I preserves the temporal order of pooled V-JEPA 2 features and interaction turns and generates dedicated interaction tokens. We evaluate on FuGIBench, a streaming-video benchmark for preparation, response, and cancellation in tool-handover assistance. The best-performing variants of EviBridge-E and EviBridge-I achieve Overall scores of 0.6448 and 0.5892, compared with 0.2178 and 0.5327 for their respective task-adapted native baselines. These gains span target handovers, long waits, and cancelled requests, demonstrating the value of predictive visual evidence for interaction decisions.
Authors
- C. Zheng (ORCID: https://orcid.org/0000-0002-7057-6609)
- Gufeng Liu (ORCID: https://orcid.org/0000-0002-8238-804X)
- 蔡昌恆
- Song Ren (ORCID: https://orcid.org/0009-0005-7373-8502)
- Luya Wang (ORCID: https://orcid.org/0000-0002-7461-174X)
- Wenxi Wu (ORCID: https://orcid.org/0009-0008-8970-0025)
- Yichen Long
- Yuan Fang
- Yuchen Wang
- Zhichen Zhao
- Zhichao Huang
- Hongsen Liu
Publication Details
- Journal
- arXiv (Cornell University)
- Published
- 2026-11-16
- DOI
- https://doi.org/10.48550/arxiv.2611.20916
- Primary Topic
- Robot Manipulation and Learning
- Type
- preprint