Thesis: Document Summarization with applications to Keyword extraction and Image Retrieval

TL;DR

This paper presents methods for document summarization with applications to keyword extraction, image retrieval, and opinion summarization, using probabilistic models, rank aggregation, and submodular functions to improve relevance and sentiment preservation.

Contribution

It introduces novel approaches combining probabilistic models, rank aggregation, and submodular functions for enhanced keyword extraction, image recommendation, and sentiment-aware summarization.

Findings

Proposed image retrieval method outperforms existing baselines.

Rank aggregation with relevance feedback improves image and text retrieval.

Submodular functions effectively balance sentiment and summarization quality.

Abstract

Automatic summarization is the process of reducing a text document in order to generate a summary that retains the most important points of the original document. In this work, we study two problems - i) summarizing a text document as set of keywords/caption, for image recommedation, ii) generating opinion summary which good mix of relevancy and sentiment with the text document. Intially, we present our work on an recommending images for enhancing a substantial amount of existing plain text news articles. We use probabilistic models and word similarity heuristics to generate captions and extract Key-phrases which are re-ranked using a rank aggregation framework with relevance feedback mechanism. We show that such rank aggregation and relevant feedback which are typically used in Tagging Documents, Text Information Retrieval also helps in improving image retrieval. These queries are fed to the Yahoo Search Engine to obtain relevant images 1. Our proposed method is observed to perform better than all existing baselines. Additonally, We propose a set of submodular functions for opinion summarization. Opinion summarization has built in it the tasks of summarization and sentiment detection. However, it is not easy to detect sentiment and simultaneously extract summary. The two tasks conflict in the sense that the demand of compression may drop sentiment bearing sentences, and the demand of sentiment detection may bring in redundant sentences. However, using submodularity we show how to strike a balance between the two requirements. Our functions generate summaries such that there is good correlation between document sentiment and summary sentiment along with good ROUGE score. We also compare the performances of the proposed submodular functions.