A probabilistic graphical model for joint answer ranking in question answering
Published 23 July 2007
Jeongwoo Ko, Eric Nyberg, Luo Si
Citations70
Generate an AI Snapshot to get a quick, structured summary of this paper.
Study Snapshot
ObjectiveStudy objective
MethodsResearch methodology
PopulationPopulation studied
Sample sizeSample sizes
OutcomesStudy outcomes here
ResultsStudy results comes here
LimitationsResearch study limitations comes here
A concise AI-generated summary of the paper will appear here once you click Generate AI Snapshot.
TL;DR
A probabilistic graphical model is applied for answer ranking in question answering which estimates the joint probability of correctness of all answer candidates, from which the probability of Correctness of an individual candidate can be inferred.
Abstract
Graphical models have been applied to various information retrieval and natural language processing tasks in the recent literature. In this paper, we apply a probabilistic graphical model for answer ranking in question answering. This model estimates the joint probability of correctness of all answer candidates, from which the probability of correctness of an individual candidate can be inferred. The joint prediction model can estimate both the correctness of individual answers as well as their correlations, which enables a list of accurate and comprehensive answers. This model was compared with a logistic regression model which directly estimates the probability of correctness of each individual answer candidate. An extensive set of empirical results based on TREC questions demonstrates the effectiveness of the joint model for answer ranking. Furthermore, we combine the joint model with the logistic regression model to improve the efficiency and accuracy of answer ranking.
Keywords
Computer Science
The MIT Press eBooksParallel Distributed Processing
15,227 Citations1986David E. Rumelhart, James L. McClelland +1 more
Language<b>WordNet: An electronic lexical database</b> . Ed. by Christiane Fellbaum. Cambridge, MA: MIT Press, 1998. Pp. xxii, 423.
11,687 Citations2000Adam Kilgarriff
The lexical database: nouns in WordNet, George A. Miller modifiers in WordNet, Katherine J. Miller a semantic network of English verbs, and applications of WordNet: building semantic concordances are presented.
Machine LearningAn Introduction to Variational Methods for Graphical Models
3,775 Citations1999Michael I. Jordan, Zoubin Ghahramani +2 more
This paper presents a tutorial introduction to the use of variational methods for inference and learning in graphical models (Bayesian networks and Markov random fields), and describes a general framework for generating variational transformations based on convex duality.
Graphical ModelsLearning and Relearning in Boltzmann Machines
1,052 Citations2001Geoffrey E. Hinton, T. J. Sejnowski
This chapter contains sections titled: Relaxation Searches, Easy and Hard learning, The Boltzmann Machine Learning Algorithm, An Example of Hard Learning, Achieving Reliable Computation with Unreliable Hardware, and an Example of the Effects of Damage.
Retrieval and novelty detection at the sentence level
281 Citations2003James Allan, Courtney Wade +1 more
This study investigates the more difficult two-part task defined by the TREC 2002 novelty track: given a topic and a group of documents relevant to that topic, find the relevant sentences from the documents, and 2) find the novel sentence from the collection of relevant sentences.
FALCON: Boosting Knowledge for Answer Engines
249 Citations2000Sanda M. Harabagiu, Dan Moldovan +7 more
FALCON, an answer engine that integrates different forms of syntactic, semantic and pragmatic knowledge for the goal of achieving better performance is discussed.
Exploiting redundancy in question answering
237 Citations2001Charles L. A. Clarke, Gordon V. Cormack +1 more
A method for arbitrary passage retrieval is applied to answer brief factual questions of the form typically asked in trivia quizzes and television game shows and it is demonstrated that answer redundancy can be used to address the second half.
A comparison of numerical optimizers for logistic regression
236 Citations2004Thomas P. Minka
This note compares eight different algorithms for computing the maximum a-posteriori parameter estimate and finds the fastest algorithms turn out to be conjugate gradient ascent and quasi-Newton algorithms, which far outstrip Iterative Scaling and its variants.
Question-answering by predictive annotation
214 Citations2000John Prager, Eric W. Brown +2 more
Predictive Annotation identifies potential answers to questions in text, annotates them accordingly and indexes them, and produces a system effective at answering natural-language fact-seeking questions posed against large document collections.
Integrating Web-based and Corpus-based Techniques for Question Answering.
58 Citations2003Boris Katz, Jimmy Lin +7 more
MIT CSAIL’s entry in the TREC Question Answering track focused on integrating Web-based techniques with more traditional strategies based on document retrieval and named-entity detection, and identified this class of techniques as the knowledge mining approach to question answering.
In question answering, two heads are better than one
57 Citations2003Jennifer Chu‐Carroll, Krzysztof Czuba +2 more
This work developed a multi-strategy and multi-source approach to question answering which is based on combining the results from different answering agents searching for answers in multiple corpora by combining results from the answering agents at the question, passage, and/or answer levels.
TREC 2002 QA at BBN: Answer Selection and Confidence Estimation.
42 Citations2002Jinxi Xu, Ana Licuanan +3 more
It is found that some simple constraints on the candidate answers can improve a pure IR-based technique for answer selection and a few simple features derived from the questionanswer pairs can be used for effective confidence estimation.
A Probabilistic Framework for Answer Selection in Question Answering
37 Citations2007Jeongwoo Ko, Luo Si +1 more
This work focuses on developing a unified framework that not only uses multiple resources for validating answer candidates, but also considers evidence of similarity among answer candidates in order to boost the ranking of the correct answer.
The University of Amsterdam at CLEF@QA 2007
35 Citations2007Christof Monz, Maarten de Rijke
The official runs of the team for the CLEF 2004 ad hoc tasks are described, including the FlexIR system as well as the approaches used for each of the tasks in which they participated.
European Conference on Artificial IntelligenceType checking in open-domain question answering
30 Citations2004Stefan Schlobach, Marius Olsthoorn +1 more
This work compares two kinds of strategies for answer type checking for open domain QA, one is redundancy-based and builds on the intuition that the amount of implicit knowledge which connects an answer to a question can be estimated by exploiting the redundancy of information available on the web, and the other is knowledge-intensive.
Defense Technical Information Center (DTIC)IBM's PIQUANT in TREC2003
27 Citations2003John Prager, Jennifer Chu‐Carroll +4 more
The analysis of the contribution of Cyc this year showed that the major limiting factor is still in the area of coverage, and the implementation of QA-by-Dossier to answer Definition questions was the major novelty.
QUALIFIER In TREC-12 QA Main Task.
26 Citations2003Hui Yang, Hang Cui +4 more
This paper tackles the factoid QA task by Event-based Question Answering by analyzing the external knowledge from pre-retrieved TREC documents, Web documents, WordNet and Ontology to discover the QA event structure, and explores the inherent associations among QA elements and then obtain the answers.
JAVELIN I and II Systems at TREC 2005
25 Citations2005Eric Nyberg, Robert Frederking +9 more
The JAVELIN team at Carnegie Mellon University submitted three question-answering runs for the TREC 2005 evaluation and the JAVelin I system was used to generate a single submission to the main track and theJAVELin II system was use to generate two submissions to the relationship track.
Text, speech and language technologyHow To Select An Answer String?
18 Citations2006Abdessamad Echihabi, Ulf Hermjakob +4 more
Three distinct approaches to solving the answer selection problem are studied: one approach uses algorithms that rely on rich knowledge bases and sophisticated syntactic/semantic processing; one approaches uses patterns that are learned in an unsupervised manner from the web, using computational biology-inspired alignment algorithms; and one approach use statistical noisy-channel algorithms similar to those used in machine translation.
Lecture notes in computer scienceQuartz: A Question Answering System for Dutch
7 Citations2007David Ahn, Valentin Jijkoun +3 more
An overview of the question classification module, the multi-dimensional markup for data storage and access, the answer processing component, and the probabilistic approach to estimating answer correctness are given.