Towards Neuro-Symbolic Video Understanding
This addresses the need for efficient frame retrieval in video data, particularly for applications like self-driving, though it is incremental as it builds on existing vision-language models.
The paper tackles the problem of long-term temporal reasoning in video understanding by decoupling semantic understanding from temporal reasoning, using vision-language models for frames and temporal logic for event evolution, achieving a 9-15% F1 score improvement on self-driving datasets.
The unprecedented surge in video data production in recent years necessitates efficient tools to extract meaningful frames from videos for downstream tasks. Long-term temporal reasoning is a key desideratum for frame retrieval systems. While state-of-the-art foundation models, like VideoLLaMA and ViCLIP, are proficient in short-term semantic understanding, they surprisingly fail at long-term reasoning across frames. A key reason for this failure is that they intertwine per-frame perception and temporal reasoning into a single deep network. Hence, decoupling but co-designing semantic understanding and temporal reasoning is essential for efficient scene identification. We propose a system that leverages vision-language models for semantic understanding of individual frames but effectively reasons about the long-term evolution of events using state machines and temporal logic (TL) formulae that inherently capture memory. Our TL-based reasoning improves the F1 score of complex event identification by 9-15% compared to benchmarks that use GPT4 for reasoning on state-of-the-art self-driving datasets such as Waymo and NuScenes.