IBM's Statistical Question Answering System - TREC-11

Ittycheriah, Abraham; Roukos, Salim

doi:10.21236/ada456310

Cited by 77 publications

(107 citation statements)

References 4 publications

Supporting

Mentioning

105

Contrasting

Unclassified

Order By: Relevance

“…Many techniques have been investigated in this area. The most successful techniques were the ones based on density [17], [20], [3](JIRS is based on density, see Figure 2 and the JIRS architecture description below) and the ones based on terms overlap [4], [9]. However, there are other works which investigated the efficiency of the PR module when the the order of the question terms is respected [2] and the possibility of using semantic information to obtain the relevant passages [16].…”

Section: The Arabic-jirs Passage Retrieval Systemmentioning

confidence: 99%

Adapting the JIRS Passage Retrieval System to the Arabic Language

Benajiba

Rosso

Soriano

2007

Lecture Notes in Computer Science

View full text Add to dashboard Cite

Abstract. The need of having a Passage Retrieval (PR) system for Arabic texts is due essentially to our aim to build an Arabic Question Answering (QA) system in our research team. We have chosen working on the PR system to be our first step to pursue our aim because being the core component and its quality will affect directly the performance of the QA system. JAVA Information Retrieval System (JIRS) is a PR QA-oriented system, multi-platform, open source and free to use. JIRS uses an n-gram model and it is language-independent. It separates language configuration files to make easier its adaptation to any language. In this paper, we report the different challenges when adapting the JIRS to the Arabic language.In order to evaluate JIRS on Arabic, we had to develop an Arabic test-bed using the multilingual CLEF QA one as guideline. We also report the results obtained in our experiments where we retrieved Arabic passages with JIRS first without any text preprocessing and second performing a prior light-stemming on the documents of the test-bed. The preliminary results show that it is possible to obtain a first Arabic passage retrieval system adapting JIRS on pre-processed text with a light-stemmer.

show abstract

Section: The Arabic-jirs Passage Retrieval Systemmentioning

confidence: 99%

Adapting the JIRS Passage Retrieval System to the Arabic Language

Benajiba

Rosso

Soriano

2007

Lecture Notes in Computer Science

View full text Add to dashboard Cite

show abstract

“…Although question classification plays a vital role in most QA systems, many factors influence the overall ability of a system to produce the correct answer to a given question. It has been shown that parallel improvements in question classification accuracy, retrieval of candidate answer, named entity tagging, and answer extraction are needed to improve the overall performance of a QA system [12].…”

Section: Overviewmentioning

confidence: 99%

“…Since it is not possible to list all such systems, we briefly describe several. Among these are IBM's TREC-9 system [12] that utilizes maximum entropy models [7]. It uses a mix of syntactic and semantic features (see Section 5).…”

Section: Related Workmentioning

confidence: 99%

Analysis of Statistical Question Classification for Fact-Based Questions

2005

View full text Add to dashboard Cite

Abstract. Question classification systems play an important role in question answering systems and can be used in a wide range of other domains. The goal of question classification is to accurately assign labels to questions based on expected answer type. Most approaches in the past have relied on matching questions against hand-crafted rules. However, rules require laborious effort to create and often suffer from being too specific. Statistical question classification methods overcome these issues by employing machine learning techniques. We empirically show that a statistical approach is robust and achieves good performance on three diverse data sets with little or no hand tuning. Furthermore, we examine the role different syntactic and semantic features have on performance. We find that semantic features tend to increase performance more than purely syntactic features. Finally, we analyze common causes of misclassification error and provide insight into ways they may be overcome.

show abstract

“…Rule 1: Question word in a chunk of length more than one (see Example (1) in Table 2). Qp = question word + headword in the same chunk…”

Section: Question Pattern Extractionmentioning

confidence: 99%

“…An automated question answering (QA) system receives a user's natural-language question and returns exact answers by analyzing the question and consulting a large text collection [1,2]. As Moldovan et al [3] pointed out, over 60% of the QA errors can be attributed to ineffective question processing, including query formulation and query expansion.…”

Section: Introductionmentioning

confidence: 99%

Web-Based Unsupervised Learning for Query Formulation in Question Answering

Wang

Liang

et al. 2005

Lecture Notes in Computer Science

View full text Add to dashboard Cite

Abstract. Converting questions to effective queries is crucial to open-domain question answering systems. In this paper, we present a web-based unsupervised learning approach for transforming a given natural-language question to an effective query. The method involves querying a search engine for Web passages that contain the answer to the question, extracting patterns that characterize fine-grained classification for answers, and linking these patterns with n-grams in answer passages. Independent evaluation on a set of questions shows that the proposed approach outperforms a naive keywordbased approach in terms of mean reciprocal rank and human effort.

show abstract

IBM's Statistical Question Answering System - TREC-11

Cited by 77 publications

References 4 publications

Adapting the JIRS Passage Retrieval System to the Arabic Language

Adapting the JIRS Passage Retrieval System to the Arabic Language

Analysis of Statistical Question Classification for Fact-Based Questions

Web-Based Unsupervised Learning for Query Formulation in Question Answering

Contact Info

Product

Resources

About