Retrieving answers from frequently asked questions pages on the web

Abstract

We address the task of answering natural language questions by using the large number of Frequently Asked Questions (FAQ) pages available on the web. The task involves three steps: (1) fetching FAQ pages from the web; (2) automatic extraction of question/answer (Q/A) pairs from the collected pages; and (3) answering users' questions by retrieving appropriate Q/A pairs. We discuss our solutions for each of the three tasks, and give detailed evaluation results on a collected corpus of about 3.6Gb of text data (293K pages, 2.8M Q/A pairs), with real users' questions sampled from a web search engine log. Specifically, we propose simple but effective methods for Q/A extraction and investigate task-specific retrieval models for answering questions. Our best model finds answers for 36% of the test questions in the top 20 results. Our overall conclusion is that FAQ pages on the web provide an excellent resource for addressing real users' information needs in a highly focused manner.

Boosting a Rule-Based Chatbot Using Statistics and User Satisfaction Ratings

“…For example, [15] reports on an attempt to answer open-domain questions asked by users on Web forums, by searching the answer in a large but limited set of FAQ QA pairs collected in a previous step. The authors use simple vector-space retrieval models over the user's question treated as a query and the FAQ question, answer, and source document indexed as fields making up the item to be returned.…”

“…The potential discrepancy between a user's actual information need and what may be inferred from its expression in a textual query is a pervasive problem in information retrieval [21]. As an example, to assess how well suited to a question the answers retrieved by their system were, the authors of [15] had raters "back-generate" a possible information need behind each question before judging the quality of the answers provided by the system. Those researchers point out that for some questions the assessors were unable to reconstruct the original information need, which means they were unable to judge the quality of the answers.…”

Generating reformulation trees for complex queries

“…Answers and Quora. Previous work [9,10,24] studied effectively finding previously answered questions that are relevant to the new question asked by the user. Different retrieval models have been proposed to calculate the similarity between questions.…”

A syntactic tree matching approach to finding similar questions in community-based qa services

“…Some recent works [15,17] used more advanced translation-based approaches to retrieve FAQ data. [11] and [10] mined the FAQ data from the Web and implemented their own retrieval systems.…”