Retrieving Passages and Finding Answers

Abstract

Retrieving topically-relevant text passages in documents has been studied many times, but finding non-factoid, multiple sentence answers to web queries is a different task that is becoming increasingly important for applications such as mobile search. As the first stage of developing retrieval models for "answer passages", we describe the process of creating a test collection of questions and multiple-sentence answers based on the TREC GOV2 queries and documents. This annotation shows that most of the description-length TREC queries do in fact have passage-level answers. We then examine the effectiveness of current passage retrieval models in terms of finding passages that contain answers. We show that the existing methods are not effective for this task, and also observe that the relative performance of these methods in retrieving answers does not correspond to their performance in retrieving relevant documents.

Query Expansion for Sentence Retrieval Using Pseudo Relevance Feedback and Word Embedding

“…Retrieval of sentences relevant to an answer is a difficult task due to short length of the target items which are more likely to suffer from vocabulary mismatch with respect to the query. We focus on the new task of answer passage retrieval for non-factoid queries [3,8]. In this paper we investigate two unsupervised query expansion (QE) methods which seek to address the query-document term mismatch issue.…”

On the Effectiveness of Contextualisation Techniques in Spoken Query Spoken Content Retrieval

“…The kernel function is commonly parametrised by a propagation parameter σ which adjusts the influence that a term occurrence has over distant positions. Among these, the Gaussian kernel exp(−(j − i) 2 /2σ 2 ) has been shown effective in previous studies [10,6,8].…”

“…The process of taking context into account when computing the relevance score of an element is known as contextualisation [3]. Various contextualisation techniques have been proven effective in XML retrieval [3], passage retrieval [6,8], and SCR [12,15] tasks. In these techniques, elements are scored depending not only on the query terms occurring within the element itself but also on those occurring in other positions within the document.…”

“…Two simple and widely adopted approaches are interpolating the scores of an element with those of its document to consider global context [12], or with those from a fixed number of surrounding elements to consider local context [15]. In contrast, techniques based on positional models (PM) [6,8] allow consideration of longer-spans of context ignoring element boundaries.…”

Learning Contextualized Document Representations for Healthcare Answer Retrieval

“…Typically, these models are trained with labeled question-answer pairs. However, it was shown that these models are not suitable for extracting local aspects from long documents, and especially not for open-ended, long answer passages [27,45,50,52]. We therefore frame our task as a passage retrieval problem, where the system's goal is to extract a concise snippet (typically 5-20 sentences) out of a large number of long documents.…”