Comparative Analysis of Word Embeddings in Assessing Semantic Similarity of Complex Sentences

Chandrasekaran, Dhivya; Mago, Vijay

doi:10.1109/access.2021.3135807

Cited by 6 publications

(2 citation statements)

References 40 publications

(32 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Furthermore, a multimodal embedding method is proposed to address complexity classification in terms of semantic, syntactic, and lexical perspectives simultaneously, replicating the research of Gargiulo et al, [34]. The dependency tree and part-ofspeech embeddings are used for the syntactic complexity classification, along with the linguistic rules, while word embeddings are used to address the semantic complexity classification, as studied by Chandrasekaran and Mago [35], and also to address the lexical complexity. Prior to the complexity classification, fundamental NLP pre-processing operations such as lower casing, stop word filtering, and tokenization is performed.…”

Section: Hybrid Personalized Text Simplification Frameworkmentioning

confidence: 98%

A Hybrid Personalized Text Simplification Framework Leveraging the Deep Learning-based Transformer Model for Dyslexic Students

Safura Adeela Sukiman,

Nor Azura Husin,

Hazlina Hamdan

et al. 2023

ARASET

View full text Add to dashboard Cite

This study proposes a hybrid personalized text simplification framework leveraging the deep learning-based Transformer model to generate simplified expository texts by addressing all sentence perspectives: semantic, syntactic, and lexical. This study targets dyslexic students due to its increasing population in the education context. Dyslexia is a learning disability characterized by reading deficiency and cognitive weakness. Thus, they need a more personalized learning experience i.e., personalized text simplification to support their classroom learning. Unfortunately, the current models of personalized text simplification can only address the syntactic and lexical perspectives of sentences, ignoring the semantic perspective. Other models employed text complexity classification at the beginning of the text simplification workflow with the intention to address the personalization element. Still, no mapping to the deficiencies of its intended users was made, and the semantic perspective of sentences remains under study. Therefore, this study is conducted to introduce hybrid methods to enhance the current personalization elements, as well as to accommodate generation of simplified expository texts at all sentence perspectives. An extensive literature was conducted using established online databases. The proposed hybrid framework is further divided into three distinct phases: Phase 1) two-phase personalization, Phase 2) multi-label text complexity classification, and Phase 3) explicit editing. It is expected that a successful implementation of the proposed hybrid personalized text simplification framework can accelerate the learning motivations of dyslexic students, hence increasing their academic achievements and reducing academic dropout rates.

show abstract

Section: Hybrid Personalized Text Simplification Frameworkmentioning

confidence: 98%

A Hybrid Personalized Text Simplification Framework Leveraging the Deep Learning-based Transformer Model for Dyslexic Students

Safura Adeela Sukiman,

Nor Azura Husin,

Hazlina Hamdan

et al. 2023

ARASET

View full text Add to dashboard Cite

show abstract

“…Hence, it is necessary to convert the text samples into numerical vectors. This conversion of raw text data into numerical values is called a word embedding technique [31]. There are two types of word embedding techniques: Frequency-based and Prediction-based.…”

Section: A Architecture Of Proposed and Related Modelsmentioning

confidence: 99%

BERT-Based Hybrid RNN Model for Multi-class Text Classification to Study the Effect of Pre-trained Word Embeddings

Shreyashree¹,

Sunagar²,

Rajarajeswari³

et al. 2022

IJACSA

View full text Add to dashboard Cite

Due to the Covid-19 pandemic which started in the year 2020, many nations had imposed lockdown to curb the spread of this virus. People have been sharing their experiences and perspectives on social media on the lockdown situation. This has given rise to increased number of tweets or posts on social media. Multi-class text classification, a method of classifying a text into one of the pre-defined categories, is one of the effective ways to analyze such data that is implemented in this paper. A Covid-19 dataset is used in this work consisting of fifteen predefined categories. This paper presents a multi-layered hybrid model, LSTM followed by GRU, to integrate the benefits of both the techniques. The advantages of word embeddings techniques like GloVe and BERT have been implemented and found that, for three epochs, the transfer learning based pre-trained BERThybrid model performs one percent better than GloVe-hybrid model but the state-of-the-art, fine-tuned BERT-base model outperforms the BERT-hybrid model by three percent, in terms of validation loss. It is expected that, over a larger number of epochs, the hybrid model might outperform the fine-tuned model.

show abstract

SiMaLSTM-SNP: novel semantic relatedness learning model preserving both Siamese networks and membrane computing

Gu,

Chen,

et al. 2023

J Supercomput

View full text Add to dashboard Cite

Comparative Analysis of Word Embeddings in Assessing Semantic Similarity of Complex Sentences

Cited by 6 publications

References 40 publications

A Hybrid Personalized Text Simplification Framework Leveraging the Deep Learning-based Transformer Model for Dyslexic Students

A Hybrid Personalized Text Simplification Framework Leveraging the Deep Learning-based Transformer Model for Dyslexic Students

BERT-Based Hybrid RNN Model for Multi-class Text Classification to Study the Effect of Pre-trained Word Embeddings

SiMaLSTM-SNP: novel semantic relatedness learning model preserving both Siamese networks and membrane computing

Contact Info

Product

Resources

About