Learning Mixtures of Submodular Shells with Application to Document Summarization

By Hui-Ching Lin, J. Bilmes. 2012

203 citations recorded by the source corpus.

We introduce a method to learn a mixture of submodular "shells" in a large-margin setting. A submodular shell is an abstract submodular function that can be instantiated with a ground set and a set of parameters to produce a submodular function. A mixture of such shells can then also be so instantiated to produce a more complex submodular function. What our algorithm learns are the mixture weights over such shells. We provide a risk bound guarantee when learning in a large-margin structured-prediction setting using a projected subgradient method when only approximate submodular optimization is possible (such as with submodular function maximization). We apply this method to the problem of multi-document summarization and produce the best results reported so far on the widely used NIST DUC-05 through DUC-07 document summarization corpora.

Subfields: Artificial Intelligence

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

Citations in the corpus

Determinantal Point Processes for Machine Learning (2012, neutral): “We also compare against the three highestscoring systems that actually competed in the DUC 2004 competition — peers 65, 104, and 35 — as well as the submodular graph-based approach recently described by Lin and Bilmes [94], which we refer to as submod1, and the improved submodular learning approach ”
Deep Metric Learning via Facility Location (2016, methodology): “This is called the facility location function, and has been widely used in data summarization and clustering [20, 34].”
Recent advances in document summarization (2017, background): “2) submodular maximization [116, 117], parameterized [118, 172], volume maximization [93, 94, 61, 207],”
From Extractive to Abstractive Summarization: A Journey (2019, background): “corpus again, Neural sentence extractor performs the best for ROUGE-2 and ROUGE-4, while the submodular[43] technique gives the best result for ROUGE-1.”
Improving the Estimation of Word Importance for News Multi-Document Summarization (2014, methodology): “Apart from these, we compare our model with two state-of-the-art systems, including the submodular approach (SUBMOD) (Lin and Bilmes, 2012) and the determinantal point process (DPP) summarizer (Kulesza and Taskar, 2012).”
Diversity and Inclusion Metrics in Subset Selection (2020, background): “Geometric or distance-based measures of diversity have also been explored within the sciences, measuring the diversity of a dataset by the dataset’s volume [1, 8, 11, 17, 20, 22, 37], variance as in PCA [30], or other measures of spread.”
Coresets for Data-efficient Training of Machine Learning Models (2019, methodology): “The facility location function has been used in various summarization applications (Lin et al., 2009; Lin & Bilmes, 2012).”
Improving the Estimation of Word Importance for News Multi-Document Summarization - Extended Technical Report (2016, methodology): “Apart from these, we compare our model with two state-of-the-art systems, including the submodular approach (SUBMOD) (Lin and Bilmes, 2012) and the determinantal point process (DPP) summarizer (Kulesza and Taskar, 2012).”