A Multi-Similarity Neural Network for Paraphrase Detection

This study introduces a multi-similarity neural network framework for paraphrase detection, an important task in natural language processing that identifies whether two sentences convey the same meaning using different expressions. The proposed method combines various similarity measures, such as string-based similarity, semantic similarity, and embedding-based similarity, with a deep learning classifier. The framework is structured as a three-phase pipeline: preprocessing, extraction of multiple similarity features, and classification through a neural network. It employs more than 168 string similarity algorithms, semantic measures derived from WordNet, and several pre-trained embedding models to compute similarity scores. These features are aggregated and supplied to a deep neural network to determine whether sentence pairs are paraphrases. The model was evaluated on the Microsoft Research Paraphrase Corpus (MSRP) using accuracy and F1-score as performance metrics. The experimental results indicate that the proposed framework achieves 81.74% accuracy and an F1 Score of 86.6%, surpassing several existing approaches. Overall, the results suggest that integrating diverse similarity measures with neural networks enhances the identification of both explicit and nuanced paraphrases, thereby supporting advancements in text analysis and plagiarism detection systems.

[1]       Mihalcea, R.; Corley, C.; Strapparava, C. Corpus-based and knowledge-based measures of text semantic similarity. In Proceedings of the AAAI, July 2006, Vol. 6,                pp. 775–780.

[2]       Hassan, S. Measuring semantic relatedness using salient encyclopedic concepts.  PhD thesis, University of North Texas, 2011.

[3]       Rus, V.; McCarthy, P.M.; Lintean, M.C.; McNamara, D.S.; Graesser, A.C. Paraphrase Identification with Lexico-Syntactic Graph Subsumption. In Proceedings of the                FLAIRS conference, May 2008, pp. 201–206.

[4]       Islam, A.; Inkpen, D. Semantic similarity of short texts.  Recent Advances in Natural Language Processing V 2009, 309, 227–236.

[5]     Milajevs, D.; Kartsaklis, D.; Sadrzadeh, M.; Purver, M. Evaluating neural word representations in tensor-based compositional settings. arXiv preprint arXiv:1408.6179,            2014.

[6]       Fernando, S.; Stevenson, M.  A semantic similarity approach to paraphrase detection.  In Proceedings of the Proceedings of the 11th annual research colloquium                of the UK special interest group for computational linguistics, March 2008, pp. 45–52.

[7]       Kozareva, Z.; Montoyo, A. Paraphrase identification on the basis of supervised machine learning techniques. In Proceedings of the International conference on                  natural language processing (in Finland). Springer, August 2006, pp. 524–533.

[8]     Qiu, L.; Kan, M.Y.; Chua, T.S. Paraphrase recognition via dissimilarity significance classification. In Proceedings of the Proceedings of the 2006 Conference on                      Empirical Methods in Natural Language Processing, July 2006, pp. 18–26.

[9]     Ul-Qayyum, Z.; Altaf, W. Paraphrase identification using semantic heuristic features. Research Journal of Applied Sciences, Engineering and Technology 2012, 4,                4894–4904.

[10]   Blacoe, W.; Lapata, M. A comparison of vector-based representations for semantic composition. In Proceedings of the Proceedings of the 2012 joint conference                on  empirical methods in natural language processing and computational natural language learning, July 2012, pp. 546–556.

[11]     Ji, Y.; Eisenstein, J. Discriminative improvements to distributional sentence similarity. In 367 Proceedings of the Proceedings of the 2013 conference on empirical              methods in natural language 368 processing, October 2013, pp. 891–896.

[12]     El Desouki, M.I.; Gomaa, W.H.; Abdalhakim, H. A hybrid model for paraphrase detection combines pros of text similarity with deep learning. Int. J. Comput. Appl                2019, 975, 8887.

[13]     Finch, A.; Hwang, Y.S.; Sumita, E. Using machine translation evaluation techniques to determine sentence-level semantic equivalence. In Proceedings of the                        Proceedings of the third international workshop on paraphrasing (IWP2005), 2005.

[14]     Wan, S.; Dras, M.; Dale, R.; Paris, C. Using dependency-based features to take the’para-farce’out of paraphrase.   In Proceedings of the Proceedings of the                           Australasian language technology workshop 2006, November 2006, pp. 131–138.

[15]     Socher, R.; Huang, E.H.; Pennington, J.; Ng, A.Y.; Manning, C.D.   Dynamic pooling and unfolding recursive autoencoders for paraphrase detection.   In                                  Proceedings  of the NIPS, December 2011, Vol. 24, pp. 801–809.

[16]     Madnani, N.; Tetreault, J.; Chodorow, M. Re-examining machine translation metrics for paraphrase identification. In Proceedings of the Proceedings of the 2012                  Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, June 2012, pp. 182–190.

[17]     He, H.; Gimpel, K.; Lin, J. Multi-perspective sentence similarity modeling with convolutional neural networks. In Proceedings of the Proceedings of the 2015                      conference on empirical methods in natural language processing, September 2015, pp. 1576–1586.

[18]    Filice, S.; Da San Martino, G.; Moschitti, A. Structural representations for learning relations between pairs of texts. In Proceedings of the Proceedings of the 53rd              Annual Meeting of the Association for Computational Linguistics and the 7th International Joint Conference on Natural Language Processing (Volume 1: Long                   Papers), July 2015, pp. 1003–1013.

[19]     Cheng, J.; Kartsaklis, D. Syntax-aware multi-sense word embeddings for deep compositional models of meaning. arXiv preprint arXiv:1508.02354 2015.

[20]     Shahmohammadi, H.; Dezfoulian, M.H.; Mansoorizadeh, M. Paraphrase detection using LSTM networks and handcrafted features. Multimedia Tools and                                  Applications 2021, 80, 6479–6492.

[21]     Kong, L.; Han, Z.; Han, Y.; Qi, H. A deep paraphrase identification model interacting semantics with syntax. Complexity 2020.

[22]     Hany, M.; Gomaa, W.H. A Hybrid Approach to Paraphrase Detection Based on Text Similarities and Machine Learning Classifiers. In Proceedings of the 2022 2nd                International Mobile, Intelligent, and Ubiquitous Computing Conference (MIUCC). IEEE, May 2022, pp. 343–348.

[23]     Reimers, N.; Gurevych, I. Sentence-bert: Sentence embeddings using siamese bert-networks. arXiv preprint arXiv:1908.10084 2019.

[24]     Reimers, N.; Gurevych, I.  Making monolingual sentence embeddings multilingual using knowledge distillation. arXiv preprint arXiv:2004.09813 2020.

[25]     Feng, F.; Yang, Y.; Cer, D.; Arivazhagan, N.; Wang, W.  Language-agnostic BERT sentence embedding. arXiv preprint arXiv:2007.01852 2020.

[26]     Bojanowski, P.; Grave, E.; Joulin, A.; Mikolov, T.  Enriching word vectors with subword information. Transactions of the association for computational linguistics                    2017, 5, 135–146.

[27]     Joulin, A.; Grave, E.; Bojanowski, P.; Mikolov, T. Bag of tricks for efficient text classification. arXiv preprint arXiv:1607.01759 2016.

[28]     Joulin, A.; Grave, E.; Bojanowski, P.; Douze, M.; Jégou, H.; Mikolov, T. Fasttext. zip: Compressing text classification models. arXiv preprint arXiv:1612.03651 2016.

[29]     El Desouki, M.I.; Gomaa, W.H. Exploring the recent trends of paraphrase detection. International Journal of Computer Applications 2019, 975.