2022
DOI: 10.1007/s00521-022-08066-8
|View full text |Cite
|
Sign up to set email alerts
|

Contextual word embeddings for tabular data search and integration

Abstract: This paper presents a new approach to retrieve and further integrate tabular datasets (collections of rows and columns) using union and join operations. In this work, both processes were carried out using a similarity measure based on contextual word embeddings, which allows finding semantically similar tables and overcome the recall problem of lexical approaches based on string similarity. This work is the first attempt to use contextual word embeddings in the whole pipeline of table search and integration, i… Show more

Help me understand this report

Search citation statements

Order By: Relevance

Paper Sections

Select...
1

Citation Types

0
1
0

Year Published

2024
2024
2024
2024

Publication Types

Select...
2

Relationship

1
1

Authors

Journals

citations
Cited by 2 publications
(1 citation statement)
references
References 33 publications
0
1
0
Order By: Relevance
“…Previous work [51] has shown the superiority of contextual word embeddings, such as BERT and RoBERTa [52], over static word embeddings like Word2vec and fastText, as well as traditional information retrieval techniques such as BM25 [53]. For this reason, this evaluation focuses on five different language models featuring diverse architectures that produce contextual word embeddings: General Text Embeddings [56]: GTE models primarily rely on the BERT framework and currently come in three sizes: large, base, and small.…”
Section: Large Language Modelsmentioning
confidence: 99%
“…Previous work [51] has shown the superiority of contextual word embeddings, such as BERT and RoBERTa [52], over static word embeddings like Word2vec and fastText, as well as traditional information retrieval techniques such as BM25 [53]. For this reason, this evaluation focuses on five different language models featuring diverse architectures that produce contextual word embeddings: General Text Embeddings [56]: GTE models primarily rely on the BERT framework and currently come in three sizes: large, base, and small.…”
Section: Large Language Modelsmentioning
confidence: 99%