Extending Word2Vec with Domain-Specific Labels

Kód výsledku v IS VaVaI

<a href="https://www.isvavai.cz/riv?ss=detail&h=RIV%2F61989100%3A27510%2F22%3A10250379" target="_blank" >RIV/61989100:27510/22:10250379 - isvavai.cz</a>

Výsledek na webu

<a href="https://annals-csis.org/Volume_30/" target="_blank" >https://annals-csis.org/Volume_30/</a>

DOI - Digital Object Identifier

—

Jazyk výsledku

angličtina

Název v původním jazyce

Extending Word2Vec with Domain-Specific Labels

Popis výsledku v původním jazyce

Choosing a proper representation of textual data isan important part of natural language processing. One option is using Word2Vec embeddings, i.e., dense vectors whose properties can to a degree capture the &quot;meaning&quot; of each word. One of the main disadvantages of Word2Vec is its inability to distinguish between antonyms. Motivated by this deficiency, this paper presents a Word2Vec extension for incorporating domain-specific labels. The goal is to improve the ability to differentiate between embeddings of words associated with different document labels or classes. This improvement is demonstrated on word embeddings derived from tweets related to a publicly traded company. Each tweet is given a label depending on whether its publication coincides with a stock price increase or decrease. The extended Word2Vec model then takes this label into account. The user can also set the weight of this label in the embedding creation process. Experiment results show that increasing this weight leads to a gradual decrease in cosine similarity between embeddings of words associated with different labels. This decrease in similarity can be interpreted as an improvement of the ability to distinguish between these words.

Název v anglickém jazyce

Extending Word2Vec with Domain-Specific Labels

Popis výsledku anglicky

Choosing a proper representation of textual data isan important part of natural language processing. One option is using Word2Vec embeddings, i.e., dense vectors whose properties can to a degree capture the &quot;meaning&quot; of each word. One of the main disadvantages of Word2Vec is its inability to distinguish between antonyms. Motivated by this deficiency, this paper presents a Word2Vec extension for incorporating domain-specific labels. The goal is to improve the ability to differentiate between embeddings of words associated with different document labels or classes. This improvement is demonstrated on word embeddings derived from tweets related to a publicly traded company. Each tweet is given a label depending on whether its publication coincides with a stock price increase or decrease. The extended Word2Vec model then takes this label into account. The user can also set the weight of this label in the embedding creation process. Experiment results show that increasing this weight leads to a gradual decrease in cosine similarity between embeddings of words associated with different labels. This decrease in similarity can be interpreted as an improvement of the ability to distinguish between these words.

Druh

D - Stať ve sborníku

CEP obor

—

OECD FORD obor

10201 - Computer sciences, information science, bioinformathics (hardware development to be 2.2, social aspect to be 5.8)

Projekt

—

Návaznosti

S - Specificky vyzkum na vysokych skolach

Rok uplatnění

2022

Kód důvěrnosti údajů

S - Úplné a pravdivé údaje o projektu nepodléhají ochraně podle zvláštních právních předpisů

Název statě ve sborníku

Annals of Computer Science and Information Systems. Volume 30

ISBN

978-83-965897-1-2

ISSN

2300-5963

e-ISSN

—

Počet stran výsledku

4

Strana od-do

157-160

Název nakladatele

Polskie Towarzystwo Informatyczne

Místo vydání

Varšava

Místo konání akce

Sofie

Datum konání akce

4. 9. 2022

Typ akce podle státní příslušnosti

WRD - Celosvětová akce

Kód UT WoS článku

000904404400022