DLS$@$CU: Sentence Similarity from Word Alignment

Sultan, Arafat; Bethard, Steven; Sumner, Tamara

doi:10.3115/v1/s14-2039

Cited by 57 publications

(21 citation statements)

References 4 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…This is because the similarities between short and grammatically simple sentences of MSRvid-12 highly depend on the syntactic roles the concepts occupy (e.g., "The man is playing with the dog" vs. "The dog is playing with the man"). Whereas the model of Sultan et al (2014), using dependency parsing between sentences can take syntactic roles of concepts into account to better align words between sentences, our simple STS metric -designed to be language-independent and resource-light -cannot.…”

Section: Resultsmentioning

confidence: 99%

“…The cross-lingual performances of the OptAlign model on the English-Spanish variants of the News-16 and MulSrc-16 datasets come reasonably close to the best performing models (Brychcín and Svoboda, 2016;Jimenez, 2016) from the SemEval 2016 STS cross-lingual evaluation. The 4-5% difference in performance (86% compared to 90% on News-16 and 77% compared to 82% on MulSrc-16) seems rather satisfying considering that OptAlign is very resource-light and language-independent, whereas the best-performing system (Brychcín and Svoboda, 2016) employs a full-blown MT system (Google translate), and a word alignment model that requires a dependency parser, a named entity recognizer, and a paraphrase database (Sultan et al, 2014).…”

Section: Resultsmentioning

confidence: 99%

“…Although there are many applications that would directly benefit from systems for detecting semantic similarity between short texts across languages (e.g., machine translation, plagiarism detection, cross-lingual retrieval), the cross-lingual setting has been considered only in the most recent edition of the SemEval STS task (Agirre et al, 2016) and only for a single language pair (English-Spanish). Although the best performing systems (Brychcín and Svoboda, 2016;Jimenez, 2016) produce the similarity scores that have 90% correlation with human scores, these systems are not truly cross-lingual because they merely use full-blown MT systems (e.g., Google translate) to translate one of the texts into another language which then allows them to apply resource-intensive monolingual alignment models, such as the one of Sultan et al (2014). These systems thus have limited applicability as they cannot be used for pairs of languages for which no robust MT model exists.…”

Section: Related Workmentioning

confidence: 99%

See 2 more Smart Citations

A resource-light method for cross-lingual semantic textual similarity

Glavaš

Franco-Salvador

Ponzetto

et al. 2018

Knowledge-Based Systems

View full text Add to dashboard Cite

Recognizing semantically similar sentences or paragraphs across languages is beneficial for many tasks, ranging from cross-lingual information retrieval and plagiarism detection to machine translation. Recently proposed methods for predicting cross-lingual semantic similarity of short texts, however, make use of tools and resources (e.g., machine translation systems, syntactic parsers or named entity recognition) that for many languages (or language pairs) do not exist. In contrast, we propose an unsupervised and a very resource-light approach for measuring semantic similarity between texts in different languages. To operate in the bilingual (or multilingual) space, we project continuous word vectors (i.e., word embeddings) from one language to the vector space of the other language via the linear translation model. We then align words according to the similarity of their vectors in the bilingual embedding space and investigate different unsupervised measures of semantic similarity exploiting bilingual embeddings and word alignments. Requiring only a limited-size set of word translation pairs between the languages, the proposed approach is applicable to virtually any pair of languages for which there exists a sufficiently large corpus, required to learn monolingual word embeddings. Experimental results on three different datasets for measuring semantic textual similarity show that our simple resource-light approach * Corresponding author Email addresses: goran@informatik.uni-mannheim.de (Goran Glavaš), mfranco@prhlt.upv.es (Marc Franco-Salvador), simone@informatik.uni-mannheim.de (Simone P. Ponzetto), prosso@dsic.upv.es (Paolo Rosso) reaches performance close to that of supervised and resource-intensive methods, displaying stability across different language pairs. Furthermore, we evaluate the proposed method on two extrinsic tasks, namely extraction of parallel sentences from comparable corpora and cross-lingual plagiarism detection, and show that it yields performance comparable to those of complex resource-intensive state-of-the-art models for the respective tasks.

show abstract

Section: Resultsmentioning

confidence: 99%

Section: Resultsmentioning

confidence: 99%

Section: Related Workmentioning

confidence: 99%

See 1 more Smart Citation

A resource-light method for cross-lingual semantic textual similarity

Glavaš

Franco-Salvador

Ponzetto

et al. 2018

Knowledge-Based Systems

View full text Add to dashboard Cite

show abstract

“…The paraphrases in it are extracted automatically from multilingual resources. It is reported useful in many other tasks such as recognizing textual entailment [31], [32], measuring the semantic similarity [33]- [35], monolingual alignment [36], [37], and natural language generation [38].…”

Section: B the Lexicon Layermentioning

confidence: 99%

Alleviating overfitting for polysemous words for word representation estimation using lexicons

Hagiwara

2017

2017 International Joint Conference on Neural Networks (IJCNN)

View full text Add to dashboard Cite

Though there are some works on improving distributed word representations using lexicons, the improper overfitting of the words that have multiple meanings is a remaining issue deteriorating the learning when lexicons are used, which needs to be solved. An alternative method is to allocate a vector per sense instead of a vector per word. However, the word representations estimated in the former way are not as easy to use as the latter one. Our previous work uses a probabilistic method to alleviate the overfitting, but it is not robust with a small corpus. In this paper, we propose a new neural network to estimate distributed word representations using a lexicon and a corpus. We add a lexicon layer in the continuous bag-of-words model and a threshold node after the output of the lexicon layer. The threshold rejects the unreliable outputs of the lexicon layer that are less likely to be the same with their inputs. In this way, it alleviates the overfitting of the polysemous words. The proposed neural network can be trained using negative sampling, which maximizing the log probabilities of target words given the context words, by distinguishing the target words from random noises. We compare the proposed neural network with the continuous bag-of-words model, the other works improving it, and the previous works estimating distributed word representations using both a lexicon and a corpus. The experimental results show that the proposed neural network is more efficient and balanced for both semantic tasks and syntactic tasks than the previous works, and robust to the size of the corpus.

show abstract

“…The overall sentence measure depends on the number of tokens, RDF triples that entail the semantic layer. In the same area, [52] combined the words meanings and phrase context in a sentence measure. The meaning words are implied by extracting words' lemma from a dictionary, whereas phrase context usage was extracted using a huge para-phrase alignment database [53].…”

Section: Hybrid-based Methodsmentioning

confidence: 99%

Efficient Hybrid Semantic Text Similarity using Wordnet and a Corpus

Atoum¹,

Otoom²

2016

ijacsa

View full text Add to dashboard Cite

Abstract-Text similarity plays an important role in natural language processing tasks such as answering questions and summarizing text. At present, state-of-the-art text similarity algorithms rely on inefficient word pairings and/or knowledge derived from large corpora such as Wikipedia. This article evaluates previous word similarity measures on benchmark datasets and then uses a hybrid word similarity in a novel text similarity measure (TSM). The proposed TSM is based on information content and WordNet semantic relations. TSM includes exact word match, the length of both sentences in a pair, and the maximum similarity between one word and the compared text. Compared with other well-known measures, results of TSM are surpassing or comparable with the best algorithms in the literature.

show abstract

DLS$@$CU: Sentence Similarity from Word Alignment

Cited by 57 publications

References 4 publications

A resource-light method for cross-lingual semantic textual similarity

A resource-light method for cross-lingual semantic textual similarity

Alleviating overfitting for polysemous words for word representation estimation using lexicons

Efficient Hybrid Semantic Text Similarity using Wordnet and a Corpus

Contact Info

Product

Resources

About