Deep Submodular Networks for Extractive Data Summarization

TL;DR

This paper introduces Deep Submodular Networks (DSN), a novel end-to-end learning framework that models complex features for improved extractive data summarization, outperforming existing methods in image collection tasks.

Contribution

The paper presents DSN, enabling learning of complex features for summarization, surpassing simple mixture models and matching or exceeding state-of-the-art performance with fewer components.

Findings

DSN outperforms simple mixture models using off-the-shelf features.

Four submodular functions in DSN perform comparably to 594 handcrafted components.

DSN achieves significant improvements in image collection summarization.

Abstract

Deep Models are increasingly becoming prevalent in summarization problems (e.g. document, video and images) due to their ability to learn complex feature interactions and representations. However, they do not model characteristics such as diversity, representation, and coverage, which are also very important for summarization tasks. On the other hand, submodular functions naturally model these characteristics because of their diminishing returns property. Most approaches for modelling and learning submodular functions rely on very simple models, such as weighted mixtures of submodular functions. Unfortunately, these models only learn the relative importance of the different submodular functions (such as diversity, representation or importance), but cannot learn more complex feature representations, which are often required for state-of-the-art performance. We propose Deep Submodular Networks (DSN), an end-to-end learning framework that facilitates the learning of more complex features and richer functions, crafted for better modelling of all aspects of summarization. The DSN framework can be used to learn features appropriate for summarization from scratch. We demonstrate the utility of DSNs on both generic and query focused image-collection summarization, and show significant improvement over the state-of-the-art. In particular, we show that DSNs outperform simple mixture models using off the shelf features. Secondly, we also show that just using four submodular functions in a DSN with end-to-end learning performs comparably to the state-of-the-art mixture model with a hand-crafted set of 594 components and outperforms other methods for image collection summarization.