Diverse and Non-redundant Answer Set Extraction on Community QA based on DPPs

TL;DR

This paper introduces a novel approach for selecting diverse, non-redundant answer sets in community QA platforms using DPPs and BERT, improving over traditional ranking methods.

Contribution

It proposes a new task of answer set selection, leveraging DPPs and BERT to enhance answer diversity and relevance in community QA.

Findings

Outperforms baseline methods on a Japanese CQA dataset

Effectively measures answer importance and similarity with BERT

Demonstrates the benefit of diverse answer set extraction

Abstract

In community-based question answering (CQA) platforms, it takes time for a user to get useful information from among many answers. Although one solution is an answer ranking method, the user still needs to read through the top-ranked answers carefully. This paper proposes a new task of selecting a diverse and non-redundant answer set rather than ranking the answers. Our method is based on determinantal point processes (DPPs), and it calculates the answer importance and similarity between answers by using BERT. We built a dataset focusing on a Japanese CQA site, and the experiments on this dataset demonstrated that the proposed method outperformed several baseline methods.