Adversarial Training for Community Question Answer Selection Based on Multi-scale Matching
This addresses the challenge of selecting relevant answers in CQA websites, which is an incremental improvement over existing methods.
The paper tackles the problem of matching valuable answers to questions in community-based question answering by framing it as a binary classification task, using an adversarial training framework with multi-scale matching to address label imbalance and achieve state-of-the-art or similar performance on SemEval 2016 and 2017 datasets.
Community-based question answering (CQA) websites represent an important source of information. As a result, the problem of matching the most valuable answers to their corresponding questions has become an increasingly popular research topic. We frame this task as a binary (relevant/irrelevant) classification problem, and present an adversarial training framework to alleviate label imbalance issue. We employ a generative model to iteratively sample a subset of challenging negative samples to fool our classification model. Both models are alternatively optimized using REINFORCE algorithm. The proposed method is completely different from previous ones, where negative samples in training set are directly used or uniformly down-sampled. Further, we propose using Multi-scale Matching which explicitly inspects the correlation between words and ngrams of different levels of granularity. We evaluate the proposed method on SemEval 2016 and SemEval 2017 datasets and achieves state-of-the-art or similar performance.