Natural Language Query Refinement for Problem Resolution from Crowd-Sourced Semi-Structured Data

We study the problem of natural language query generation for decision support systems (DSS) in the problem resolution domain. In this domain, a user has a task he is unable to accomplish (eg. bluetooth headphone not playing music), which we capture using language structures. We show how important units that define a problem can robustly and automatically be extracted from large noisy online forum data, with no labeled data or query logs. We also show how these units can be selected to reduce the number of interactions and how they can be used to generate natural language interactions for query refinement.