TV-TREES: Multimodal Entailment Trees for Neuro-Symbolic Video Reasoning

arXiv:2402.19467v1 Announce Type: cross
Abstract: It is challenging to perform question-answering over complex, multimodal content such as television clips. This is in part because current video-language models rely on single-modality reasoning, have lowered performance on long inputs, and lack interpetability. We propose TV-TREES, the first multimodal entailment tree generator. TV-TREES serves as an approach to video understanding that promotes interpretable joint-modality reasoning by producing trees of entailment relationships between simple premises directly entailed by the videos and higher-level conclusions. We …

abstract arxiv cs.ai cs.cl cs.cv current generator inputs language language models multimodal multimodal content neuro part performance question reasoning television tree trees type video