2017/10/10 by Seunghyun Yoon, Yoon, Seunghyun, Joongbo Shin +3 · 1 citation
Computer Science · #Artificial Intelligence (cs.AI) #Computation and Language (cs.CL) #Expert finding and Q&A systems #FOS: Computer and information sciences #Natural Language Processing Techniques #Topic Modeling
paper · pdf · doi:10.48550/arxiv.1710.03430
openalex publication_date 2017/10/10 · openalex created_date 2022/10/04 · openalex updated_date 2026/07/28
In this paper, we propose a novel end-to-end neural architecture for ranking\ncandidate answers, that adapts a hierarchical recurrent neural network and a\nlatent topic clustering module. With our proposed model, a text is encoded to a\nvector representation from an word-level to a chunk-level to effectively\ncapture the entire meaning. In particular, by adapting the hierarchical\nstructure, our model shows very small performance degradations in longer text\ncomprehension while other state-of-the-art recurrent neural network models\nsuffer from it. Additionally, the latent topic clustering module extracts\nsemantic information from target samples. This clustering module is useful for\nany text related tasks by allowing each data sample to find its nearest topic\ncluster, thus helping the neural network model analyze the entire data. We\nevaluate our models on the Ubuntu Dialogue Corpus and consumer electronic\ndomain question answering dataset, which is related to Samsung products. The\nproposed model shows state-of-the-art results for ranking question-answer\npairs.\n