Weiwei: A Simple Unsupervised Latent Semantics based Approach for Sentence Similarity
Published 7 June 2012
Weiwei Guo, Mona Diab
Citations22
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 simple unsupervised latent semantics based approach, Weighted Textual Matrix Factorization that only exploits bag-of-words features, can outperform most systems for this task.
Abstract
The Semantic Textual Similarity (STS) shared task (Agirre et al., 2012) computes the degree of semantic equivalence between two sentences. 1 We show that a simple unsupervised latent semantics based approach, Weighted Textual Matrix Factorization that only exploits bag-of-words features, can outperform most systems for this task. The key to the approach is to carefully handle missing words that are not in the sentence, and thus rendering it superior to Latent Semantic Analysis (LSA) and Latent Dirichlet Allocation (LDA). Our system ranks 20 out of 89 systems according to the official evaluation metric for the task, Pearson correlation, and it ranks 10/89 and 19/89 in the other two evaluation metrics employed by the organizers. 1
Keywords
Computer Science
Journal of Machine Learning ResearchLatent dirichlet allocation
27,049 Citations2003David M. Blei, Andrew Y. Ng +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.
Discourse ProcessesAn introduction to latent semantic analysis
4,817 Citations1998Thomas K. Landauer, Peter W. Foltz +1 more
The adequacy of LSA's reflection of human knowledge has been established in a variety of ways, for example, its scores overlap those of humans on standard vocabulary and subject matter tests; it mimics human word sorting and category judgments; it simulates word‐word and passage‐word lexical priming data.
Corpus-based and knowledge-based measures of text semantic similarity
1,189 Citations2006Rada Mihalcea, Courtney D. Corley +1 more
This paper shows that the semantic similarity method out-performs methods based on simple lexical matching, resulting in up to 13% error rate reduction with respect to the traditional vector-based similarity metric.
IEEE Transactions on Knowledge and Data EngineeringSentence similarity based on semantic nets and corpus statistics
821 Citations2006Yuhua Li, David McLean +3 more
Experiments demonstrate that the proposed method provides a similarity measure that shows a significant correlation to human intuition and can be used in a variety of applications that involve text knowledge representation and discovery.
Collecting Highly Parallel Data for Paraphrase Evaluation
807 Citations2011David Chen, William B. Dolan
A novel data collection framework is presented that produces highly parallel text data relatively inexpensively and on a large scale that allows for simple n-gram comparisons to measure both the semantic adequacy and lexical dissimilarity of paraphrase candidates.
Unsupervised construction of large paraphrase corpora
740 Citations2004Bill Dolan, Chris Quirk +1 more
Investigation of unsupervised techniques for acquiring monolingual sentence-level paraphrases from a corpus of temporally and topically clustered news articles collected from thousands of web-based news sources shows that edit distance data is cleaner and more easily-aligned than the heuristic data.
Proceedings of the International AAAI Conference on Web and Social MediaCharacterizing Microblogs with Topic Models
724 Citations2010Daniel Ramage, Susan Dumais +1 more
A scalable implementation of a partially supervised learning model (Labeled LDA) that maps the content of the Twitter feed into dimensions that correspond roughly to substance, style, status, and social characteristics of posts is presented.
Extended gloss overlaps as a measure of semantic relatedness
719 Citations2003Satanjeev Banerjee, Ted Pedersen
A new measure of semantic relatedness between concepts that is based on the number of shared words (overlaps) in their definitions (glosses) and reasonably correlates to human judgments is presented.
International Conference on Machine LearningWeighted low-rank approximations
702 Citations2003Nathan Srebro, Tommi Jaakkola
This work provides a simple and efficient algorithm for solving weighted low-rank approximation problems, which, unlike their unweighted version, do not admit a closed-form solution in general.
SemEval-2012 Task 6: A Pilot on Semantic Textual Similarity
679 Citations2012Eneko Agirre, Daniel Cer +2 more
The results of the STS pilot task in Semeval open an exciting way ahead, although there are still open issues, specially the evaluation metric.
ACM Transactions on Knowledge Discovery from DataSemantic text similarity using corpus-based word similarity and string similarity
485 Citations2008Aminul Islam, Diana Inkpen
A method for measuring the semantic similarity of texts using a corpus-based measure of semantic word similarity and a normalized and modified version of the Longest Common Subsequence string matching algorithm is presented.
Training and testing of recommender systems on data missing not at random
355 Citations2010Harald Steck
It is shown that the absence of ratings carries useful information for improving the top-k hit rate concerning all items, a natural accuracy measure for recommendations, and two performance measures can be estimated, under mild assumptions, without bias from data even when ratings are missing not at random (MNAR).
Paraphrasing for automatic evaluation
221 Citations2006David Kauchak, Regina Barzilay
It is shown that the use of a paraphrased synthetic reference refines the accuracy of automatic evaluation and there is a strong connection between the quality of automatic paraphrases as judged by humans and their contribution to automatic evaluation.
Automatic evaluation of text coherence: models and representations
195 Citations2005Mirella Lapata, Regina Barzilay
A fully-automatic, linguistically rich model of local coherence that correlates with human judgments is introduced and demonstrates that certain models capture complementary aspects of coherence and thus can be combined to improve performance.
Journal of Artificial Intelligence ResearchText Relatedness Based on a Word Thesaurus
147 Citations2010George Tsatsaronis, Iraklis Varlamis +1 more
Experimental evaluation shows that the proposed method outperforms every lexicon-based method of semantic relatedness in the selected tasks and the used data sets, and competes well against corpus-based and hybrid approaches.
Modeling Sentences in the Latent Space
119 Citations2012Weiwei Guo, Mona Diab
By carefully handling words that are not in the sentences (missing words), a reliable latent variable model on sentences is trained and a new evaluation framework for sentence similarity is proposed: Concept Definition Retrieval.
Lecture notes in computer scienceA Comparative Study of Two Short Text Semantic Similarity Measures
77 Citations2008James O’Shea, Zuhair Bandar +2 more
A comparative study of STASIS and LSA is described, which shows measures of semantic similarity can be applied to short texts for use in Conversational Agents (CAs), and a benchmark data set of 65 sentence pairs with human-derived similarity ratings is presented.
Word Sense Disambiguation-based Sentence Similarity
36 Citations2010ChukFong Ho, Masrah Azrifah Azmi Murad +2 more
A method which computes the similarity between two sentences based on a comparison of their actual meanings is presented, achieved by transforming an existing most-outstanding corpus- based measure into a knowledge-based measure, which is then integrated with word sense disambiguation.
Sentence Similarity based on Relevance
23 Citations2008Feng Jin, Yiming Zhou +1 more
A novel method is proposed to estimate the sentence similarity with consideration of direct relevance and indirect relevance between sentence pairs with insights into the strengths and weakness of the proposed method.
Learning the Latent Semantics of a Concept from its Definition
6 Citations2012Weiwei Guo, Mona Diab
This paper learns low-dimensional latent semantic vectors of concept definitions to construct a more robust sense similarity measure wmfvec, achieving results comparable to state of the art WSD systems.