Macquarie University at BioASQ 6b: Deep learning and deep reinforcement learning for query-based multi-document summarisation

TL;DR

This paper presents Macquarie University's approach to query-based multi-document summarisation in BioASQ 6b, utilizing deep learning and reinforcement learning to extract ideal answers from biomedical documents.

Contribution

It introduces a deep learning model with LSTM-based features and a reinforcement learning prototype for global policy training in biomedical answer extraction.

Findings

Deep learning model achieved competitive performance.

Reinforcement learning prototype demonstrated feasibility.

Features based on LSTM, similarity, and position were effective.

Abstract

This paper describes Macquarie University's contribution to the BioASQ Challenge (BioASQ 6b, Phase B). We focused on the extraction of the ideal answers, and the task was approached as an instance of query-based multi-document summarisation. In particular, this paper focuses on the experiments related to the deep learning and reinforcement learning approaches used in the submitted runs. The best run used a deep learning model under a regression-based framework. The deep learning architecture used features derived from the output of LSTM chains on word embeddings, plus features based on similarity with the query, and sentence position. The reinforcement learning approach was a proof-of-concept prototype that trained a global policy using REINFORCE. The global policy was implemented as a neural network that used $tf.idf$ features encoding the candidate sentence, question, and context.