Learning To Recognize Procedural Activities with Distant Supervision

TL;DR

This paper introduces a method for classifying complex, multi-step activities in long videos by automatically identifying steps through distant supervision from a textual knowledge base, improving generalization across various tasks.

Contribution

The paper presents a novel approach that leverages distant supervision from wikiHow to automatically label steps in instructional videos, enabling training without manual annotations.

Findings

Models trained with automatically-labeled steps outperform baselines.

The approach generalizes well to multiple downstream tasks.

Automatic step identification improves activity recognition accuracy.

Abstract

In this paper we consider the problem of classifying fine-grained, multi-step activities (e.g., cooking different recipes, making disparate home improvements, creating various forms of arts and crafts) from long videos spanning up to several minutes. Accurately categorizing these activities requires not only recognizing the individual steps that compose the task but also capturing their temporal dependencies. This problem is dramatically different from traditional action classification, where models are typically optimized on videos that span only a few seconds and that are manually trimmed to contain simple atomic actions. While step annotations could enable the training of models to recognize the individual steps of procedural activities, existing large-scale datasets in this area do not include such segment labels due to the prohibitive cost of manually annotating temporal boundaries in long videos. To address this issue, we propose to automatically identify steps in instructional videos by leveraging the distant supervision of a textual knowledge base (wikiHow) that includes detailed descriptions of the steps needed for the execution of a wide variety of complex activities. Our method uses a language model to match noisy, automatically-transcribed speech from the video to step descriptions in the knowledge base. We demonstrate that video models trained to recognize these automatically-labeled steps (without manual supervision) yield a representation that achieves superior generalization performance on four downstream tasks: recognition of procedural activities, step classification, step forecasting and egocentric video classification.