Linear transformations for cross-lingual semantic textual similarity

Brychcín, Tomáš

doi:10.1016/j.knosys.2019.06.027

Cited by 21 publications

(27 citation statements)

References 25 publications

(53 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The global post-processing techniques for semantic spaces used by Brychcín (2020) consist of two steps column-wise mean centering and word vector normalization to unit vectors. This guarantees that all word pairs in the dictionary contribute equally to the optimization criteria of the linear transformation.…”

Section: Methodsmentioning

confidence: 99%

Evaluation Datasets for Cross-lingual Semantic Textual Similarity

Hercig¹,

Král²

2021

Proceedings of the Conference Recent Advances in Natural Language Processing - Deep Learning for Natural Language Processing Me

View full text Add to dashboard Cite

Section: Methodsmentioning

confidence: 99%

Evaluation Datasets for Cross-lingual Semantic Textual Similarity

Hercig¹,

Král²

2021

Proceedings of the Conference Recent Advances in Natural Language Processing - Deep Learning for Natural Language Processing Me

View full text Add to dashboard Cite

“…Adouane et al [32] introduced an LSTM-based neural model to detect the binary similarity label. Brychcín [33] introduced a new transformation technique to project monolingual semantic spaces using a bilingual dictionary. Lenz et al [34] proposed new supervised and unsupervised measures in the context of a graph-based similarity for argument graphs.…”

Section: Related Workmentioning

confidence: 99%

Leveraging Grammatical Roles for Measuring Semantic Similarity Between Texts

et al. 2021

View full text Add to dashboard Cite

Semantic similarity between texts can be defined based on their meaning. Assessing the textual similarity is a prerequisite in almost all applications in the field of language processing and information retrieval. However, the diversity in the sentence structure makes it formidable to estimate the similarity. Some sentences pairs are lexicographically similar but semantically dissimilar. That is why the trivial lexical overlapping is not enough for measuring the similarity. To attain the semanticity of sentences, the context of the words and the structure of the sentence should be considered. In this paper, we propose a new method for capturing the semantic similarity between sentences based on their grammatical roles through word semantics. First, the sentences are divided grammatically into different parts where each part is considered as a grammatical role. Then multiple new measures are introduced to estimate the role-based similarity exploiting word semantics considering the sentence structure. The proposed similarity measures focus on inter-role and intra-role similarity between the sentence-pair. The word-level semantic information is extracted from a pre-trained word-embedding model. The performance of the proposed method was verified by conducting a wide range of experiments on the SemEval STS dataset. The experimental results indicated the effectiveness of the proposed method in terms of different standard evaluation metrics and outperformed some known related works.

show abstract

“…The paradigmatic parameterization of English-language digital discourse will take place within the quadronymic approach (according to Descartes Square), so it seems possible to locate the Descartes Square on the graph along the abscissa (y) and ordinate (x) axes (Glavaš et al, 2018;Brychcín, 2020). The abscissa axis (y) will denote the Potential Discursive Impulse, in other words, constitutive features (features of discourse institutionality and type) and categories (genre-stylistic, substantive, formal-structural), which will be considered as potentials of the discourse (Carpenter, 1992) The square abcd is a graphical representation of English-language digital discourse.…”

Section: Quadronymic Potential Of Discursive Impulse and English-language Digital Discourse Responsementioning

confidence: 99%

Quadronimic interpretation of English-language digital discourse

Lazebna¹

2021

lingcure

View full text Add to dashboard Cite

The paper focuses on the English-language digital discourse (EDD) and considers it in terms of Descartes’ square. Despite the incongruence of the medieval philosophical paradigm with the modern stage of posthumanitarian studies, it is the quadronymic and clearly delineated frame, “perspicuae veritates” (according to R. Descartes), which can balance a diverse phenomenon of digital discourse. There are three main aspects of the study considered further: English-language digital discourse and its quadronymic potential, which is interpreted in three main perspectives: in terms of discursive impulse and discourse response; within hyper- / hypodynamic levels of English-language digital discourse and the English-language digital textual space (EDTS). Quadrominic interpretation of English-language digital discourse results in the following findings: symmetry of discourse construction; logic and sequence of discourse representation and its parameters; slot representation of discourse paradigmatic parameters; identification of interdependencies, conditions and parameters of discourse within quadrants.

show abstract

Linear transformations for cross-lingual semantic textual similarity

Cited by 21 publications

References 25 publications

Evaluation Datasets for Cross-lingual Semantic Textual Similarity

Evaluation Datasets for Cross-lingual Semantic Textual Similarity

Leveraging Grammatical Roles for Measuring Semantic Similarity Between Texts

Quadronimic interpretation of English-language digital discourse

Contact Info

Product

Resources

About