Towards a Universal Method for Measuring Semantic Textual Similarity
ثبت نشده
چکیده
Semantic textual similarity (STS) measures the degree to which two texts share the same meaning. In Natural Language Processing, STS touches many different aspects, from thesaurus generation to machine translation. However, methods for measuring STS have often been developed only for very specific types of texts, such as for comparing two words or for comparing paragraphs. This disjoint approach creates the predicament where different similarity methods have to be used depending on the type of text being compared.
منابع مشابه
CFILT-CORE: Semantic Textual Similarity using Universal Networking Language
This paper describes the system that was submitted in the *SEM 2013 Semantic Textual Similarity shared task. The task aims to find the similarity score between a pair of sentences. We describe a Universal Networking Language (UNL) based semantic extraction system for measuring the semantic similarity. Our approach combines syntactic and word level similarity measures along with the UNL based se...
متن کاملCFILT-CORE: Finding Semantic Textual Similarity using UNL
Semantic Textual Similarity is the task of finding the degree of similarity between a pair of sentences through semantics extraction. This is motivated by the fact that syntactically diverse sentences often convey the same meaning. This paper describes the approach that was used in the *SEM Shared Task 2013. The approach combines semantic, syntactic and lexical similarity measures for finding s...
متن کاملjanardhan: Semantic Textual Similarity using Universal Networking Language graph matching
Sentences that are syntactically quite different can often have similar or same meaning. The SemEval 2012 task of Semantic Textual Similarity aims at finding the semantic similarity between two sentences. The semantic representation of Universal Networking Language (UNL), represents only the inherent meaning in a sentence without any syntactic details. Thus, comparing the UNL graphs of two sent...
متن کاملMeasuring Semantic Similarity for Bengali Tweets Using WordNet
Similarity between natural language texts, sentences in terms of meaning, known as textual entailment, is a generic problem in the area of computational linguistics. In the last few years researchers worked on various aspects of textual entailment problem, but mostly restricted to English language. Here in this paper we present a method for measuring the semantic similarity of Bengali tweets us...
متن کاملMeasuring Semantic Similarity in Short Texts through Greedy Pairing and Word Semantics
We propose in this paper a greedy method to the problem of measuring semantic similarity between short texts. Our method is based on the principle of compositionality which states that the overall meaning of a sentence can be captured by summing up the meaning of its parts, i.e. the meanings of words in our case. Based on this principle, we extend wordto-word semantic similarity metrics to quan...
متن کامل