Video Captioning via Hierarchical Reinforcement Learning

TL;DR

This paper introduces a hierarchical reinforcement learning framework for detailed video captioning, effectively capturing multiple fine-grained actions and surpassing existing methods on large-scale and standard datasets.

Contribution

It proposes a novel hierarchical RL approach with high-level and low-level modules for improved fine-grained video captioning.

Findings

Outperforms all baselines on a new large-scale dataset

Achieves state-of-the-art results on MSR-VTT dataset

Effectively captures multiple fine-grained actions in videos

Abstract

Video captioning is the task of automatically generating a textual description of the actions in a video. Although previous work (e.g. sequence-to-sequence model) has shown promising results in abstracting a coarse description of a short video, it is still very challenging to caption a video containing multiple fine-grained actions with a detailed description. This paper aims to address the challenge by proposing a novel hierarchical reinforcement learning framework for video captioning, where a high-level Manager module learns to design sub-goals and a low-level Worker module recognizes the primitive actions to fulfill the sub-goal. With this compositional framework to reinforce video captioning at different levels, our approach significantly outperforms all the baseline methods on a newly introduced large-scale dataset for fine-grained video captioning. Furthermore, our non-ensemble model has already achieved the state-of-the-art results on the widely-used MSR-VTT dataset.