A Benchmark for Structured Procedural Knowledge Extraction from Cooking Videos
This work addresses the need for finer-grained evaluation of multimodal models in procedural learning, though it is incremental as it builds on existing video captioning and knowledge extraction tasks.
The authors tackled the problem of extracting structured procedural knowledge from cooking videos by proposing a new benchmark that requires models to produce interpretable verb-argument tuples, resulting in a manually annotated resource of 356 videos and 15,523 annotations, with analysis showing that standard approaches perform poorly on this task.
Watching instructional videos are often used to learn about procedures. Video captioning is one way of automatically collecting such knowledge. However, it provides only an indirect, overall evaluation of multimodal models with no finer-grained quantitative measure of what they have learned. We propose instead, a benchmark of structured procedural knowledge extracted from cooking videos. This work is complementary to existing tasks, but requires models to produce interpretable structured knowledge in the form of verb-argument tuples. Our manually annotated open-vocabulary resource includes 356 instructional cooking videos and 15,523 video clip/sentence-level annotations. Our analysis shows that the proposed task is challenging and standard modeling approaches like unsupervised segmentation, semantic role labeling, and visual action detection perform poorly when forced to predict every action of a procedure in a structured form.