Procedural Bridging for One-Image Temporal Reasoning

TL;DR: From a single image, generate a short video-of-thought that interpolates plausible steps before and after, then answer questions—like imagining the “recipe steps” around a cooking snapshot. A small-scale study on TemporalCook and Daily-PP (VG-TVP) evaluates whether these “bridged” videos improve procedural QA.

Research Question: Can text↔video bridging and procedural planning convert static images into useful “micro-simulations” that unlock temporal reasoning?

Hypothesis: Visually grounded text plans (VG-TVP) that drive short video-of-thought bridges will outperform text-only retrieval baselines on procedural inference from a single image.

Experiment Plan: - Setup: Use VG-TVP’s V2T and T2V bridges to generate stepwise textual plans and short video segments around a single image; run QA over the generated micro-simulation.

• Data/Measures: TemporalCook; human preference on temporal coherence; QA accuracy vs. RAG-only baselines.

References: ['Zarei, M., Akkasi, A., Hassan, M., & Komeili, M. (2025). TemporalCook: Benchmarking Temporal and Procedural Reasoning in Multimodal Large Language Models. Proceedings of the 1st International Workshop on MLLM for Unified Comprehension and Generation.', 'Ilaslan, M., Koksal, A., Lin, K. Q., Satar, B., Shou, M. Z., & Xu, Q. (2024). VG-TVP: Multimodal Procedural Planning via Visually Grounded Text-Video Prompting. AAAI Conference on Artificial Intelligence.', 'Tong, J., Mou, Y., Li, H., Li, M., Yang, Y., Zhang, M., Chen, Q., Liang, T., Hu, X., Zheng, Y., Chen, X., Zhao, J., Huang, X., & Qiu, X. (2025). Thinking with Video: Video Generation as a Promising Multimodal Reasoning Paradigm.']