Microsoft AI Challenge India 2018: Learning to Rank Passages for Web Question Answering with Deep Attention Networks

2019/06/14 by Chaitanya Sai Alaparthi, Alaparthi, Chaitanya Sai
Computer Science · #Topic Modeling #Multimodal Machine Learning Applications #Expert finding and Q&A systems

paper · pdf · doi:10.48550/arxiv.1906.06056

Abstract

This paper describes our system for The Microsoft AI Challenge India 2018: Ranking Passages for Web Question Answering. The system uses the biLSTM network with co-attention mechanism between query and passage representations. Additionally, we use self attention on embeddings to increase the lexical coverage by allowing the system to take union over different embeddings. We also incorporate hand-crafted features to improve the system performance. Our system achieved a Mean Reciprocal Rank (MRR) of 0.67 on eval-1 dataset.

Citations

• On the difficulty of training Recurrent Neural Networks
• Pointer Sentinel Mixture Models
• NLTK: The Natural Language Toolkit

Related

• Answering questions by learning to rank -- Learning to rank by answering questions
• Training a Ranking Function for Open-Domain Question Answering
• Learning to Focus when Ranking Answers
• A strong baseline for question relevancy ranking
• aNMM: Ranking Short Answer Texts with Attention-Based Neural Matching Model