scholarly journals Alleviating overfitting for polysemous words for word representation estimation using lexicons

2017 ◽  
Author(s):  
Yuanzhi Ke ◽  
Masafumi Hagiwara
Keyword(s):  
Word Representation ◽  
Polysemous Words

scholarly journals The Concept “Semantic Invariant" as a Cognitive Term

Discourse ◽  
2020 ◽  
Vol 6 (2) ◽  
pp. 125-133
Author(s):  
S. A. Pesina
Keyword(s):  
Native Speakers ◽  
Semantic Network ◽  
Semantic Networks ◽  
Semantic Space ◽  
Multiple Use ◽  
Word Structure ◽  
Word Representation ◽  
Characteristic Features ◽  
Language Environment ◽  
Polysemous Words

Introduction. The purpose of the paper is to prove the semantic integrity of the meanings of polysemous words based on the use of the concepts “semantic network“ and “lexical invariant“. The relevance of this study lies in the lack of sufficient support of semantic integrity of lexemes, since there are contradictory linguistic data in favor of the so-called list theory of word representation in the mental lexicon. We put forward the hypothesis of network structuring of word meanings based on the “lexical invariant“, understood as a set of dominant basic semantic components, which underlie contextual meanings in one of their configurations. We pointed out the characteristic features and advantages of the network model of polysemous words as an open system capable of building an unlimited number of connections. The study is also concerned with the determination of the way the meanings are mapped in the semantic space of the lexicon by means of revealing the content of the semantic structures of the polysemous words.Methodology and sources. The paper presents an empirical invariant-component method of analyzing a polysemous word “key“.Results and discussion. The main results of the study indicate that there are numerous interconnected semantic networks of polysemous words in the lexicon. They function as multi-level configurations of meanings, which are cemented by dominant invariant meanings. Invariant meanings are eventually formed as a result of multiple use of all metaphorical meanings clusters, which allow native speakers to effectively navigate in the surrounding language environment.Conclusion. The results of our study have proved the idea that native speakers do not need detailed information (numerous semantic components) to interpret this or that meaning. The number of components they use in everyday communication is limited but sufficient for general understanding of the words and texts contents. The research results enabled us to draw a conclusion that in order to preserve the semantic integrity of a word structure, the number of meanings must remain within the limits of the invariant semantic components. The invariant lexical components form clusters of integral and differential types that combined in different configurations form a word desired meaning.

scholarly journals Bilingual Distributed Word Representations from Document-Aligned Comparable Data

2016 ◽  
Vol 55 ◽  
pp. 953-994 ◽  
Author(s):  
Ivan Vulić ◽  
Marie-Francine Moens
Keyword(s):  
Topic Modeling ◽  
Representation Learning ◽  
Local Context ◽  
Comparable Data ◽  
Lexical Resources ◽  
Probabilistic Topic Modeling ◽  
Syntactic Information ◽  
Word Representation ◽  
Previous State ◽  
Polysemous Words

We propose a new model for learning bilingual word representations from non-parallel document-aligned data. Following the recent advances in word representation learning, our model learns dense real-valued word vectors, that is, bilingual word embeddings (BWEs). Unlike prior work on inducing BWEs which heavily relied on parallel sentence-aligned corpora and/or readily available translation resources such as dictionaries, the article reveals that BWEs may be learned solely on the basis of document-aligned comparable data without any additional lexical resources nor syntactic information. We present a comparison of our approach with previous state-of-the-art models for learning bilingual word representations from comparable data that rely on the framework of multilingual probabilistic topic modeling (MuPTM), as well as with distributional local context-counting models. We demonstrate the utility of the induced BWEs in two semantic tasks: (1) bilingual lexicon extraction, (2) suggesting word translations in context for polysemous words. Our simple yet effective BWE-based models significantly outperform the MuPTM-based and context-counting representation models from comparable data as well as prior BWE-based models, and acquire the best reported results on both tasks for all three tested language pairs.

scholarly journals Unconscious semantic processing of polysemous words is not automatic

2016 ◽  
Vol 2016 (1) ◽  
Author(s):  
Benjamin Rohaut ◽  
F.-Xavier Alario ◽  
Jacqueline Meadow ◽  
Laurent Cohen ◽  
Lionel Naccache
Keyword(s):  
Semantic Processing ◽  
Polysemous Words

scholarly journals Problems of defining legal concepts in laws and other normative legal acts

Legal Ukraine ◽  
2020 ◽  
pp. 30-41
Author(s):  
Volodymyr Klochkov
Keyword(s):  
The State ◽  
Legal Concept ◽  
Legal Norms ◽  
Legal Nature ◽  
State And Society ◽  
Word Representation ◽  
The Law ◽  
Definition Of ◽  
Legal Concepts ◽  
Laws And Regulations

Each legal concept has not only content (content), but also a form. The form requires compliance with the rules for the definition and construction of concepts. Improving legal terminology is impossible without deep development and observance of the rules of analysis and the precise construction of the conceptual apparatus. Gaps in legislation and regulations are derived from inaccuracy, lack of clarity and simplicity of conceptual constructions. The inconsistency of certain legal norms found in various laws and regulatory legal acts, the inconsistency of norms with the prevailing realities of legal life in the state and society impede the fulfillment by state authorities, including law enforcement, of their functional responsibilities. Mistakes made in the preparation of draft laws and regulations, methodological recommendations mainly boil down to the violation of the requirements of the unified laws of logic: the law of identity, the law of contradiction, the law of the excluded third and the law of sufficient reason. The use of inappropriate terminology causes complications in the application of legal norms. The Constitution of Ukraine assigns to the prosecutor the function of representing the interests of a citizen or state in court in cases specified by law. The term representation is not exactly chosen. The word "representation" means: the performance of the duties of a representative; an institution representing the interests of someone; elections, as well as the law, the procedure for the election of representatives to any bodies; representation is a legal relationship in which one party (representative) is obligated or entitled to make a transaction on behalf of the other party that it represents; representation means activity on behalf of someone, on behalf of a person. By its legal nature, a representative can only be authorized for transactions that the person he represents is entitled to carry out. The representation of the prosecutor's office in court is specific, since this body does not need instructions, contracts or other documents. The prosecutor or his deputy should act not on behalf of someone, but on behalf of the state in favor of the person and citizen, state or society, within the limits established by law. In the legislation there is a conflict (conflict) in the law regarding the term «representation». To eliminate such a conflict, it is necessary to amend the Law of Ukraine «On the Prosecutor's Office». Key words: definition of concepts, laws and regulations, accuracy, clarity, brevity of terms.

Sign in / Sign up

Export Citation Format

Share Document