Arabic Aspect Extraction Based on Stacked Contextualized Embedding With Deep Learning

Fadel, Arwa S.; Saleh, Mostafa; Abulnaja, O.A.

doi:10.1109/access.2022.3159252

Cited by 22 publications

(12 citation statements)

References 58 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Note none of the previous models for Arabic worked on multitask learning, so the comparisons were with a single task. As shown in Table 7, AR-LCF-ATEPC-Fusion achieved the best results with an F1 score of 75.94% for the ATE task, thereby outperforming all comparison models (except our previous single-task ATE method [75]). For the APC task, AR-LCF-ATEPC-Fusion outperformed all comparison models with an accuracy of 91.5% and an F1 score of 76.74%, improving the accuracy by 2%.…”

Section: Performance Of Proposed Multi-task Model On the Original Dat...mentioning

confidence: 85%

“…To validate the effectiveness of multi-task model, we compared the best multi-task model (AR-LCF-ATEPC-Fusion) with state-of-the-art Deep-based and transformer-based approaches that used the same benchmark dataset: RNN-BiLSTM-CRF [69], BiGRU [70], attention mechanism with neural network [71], BERT [72], and Bert-Flair-BiLSTM/BiGRU-CRF [75], Sequence to Sequence mode for preprocessing and BERT for classification (Seq-seq BERT) [76] and BERT with liner layer (Bert-linerpair) [77]. The results demonstrated that LCF-ATEPC model outperformed other comparable models.…”

Section: Performance Of Proposed Multi-task Model On the Original Dat...mentioning

confidence: 99%

See 1 more Smart Citation

Multi-Task Learning Model with Data Augmentation for Arabic Aspect-Based Sentiment Analysis

Fadel¹,

Abulnaja²,

Saleh³

2023

Computers, Materials &Amp; Continua

View full text Add to dashboard Cite

Aspect-based sentiment analysis (ABSA) is a fine-grained process. Its fundamental subtasks are aspect term extraction (ATE) and aspect polarity classification (APC), and these subtasks are dependent and closely related. However, most existing works on Arabic ABSA content separately address them, assume that aspect terms are preidentified, or use a pipeline model. Pipeline solutions design different models for each task, and the output from the ATE model is used as the input to the APC model, which may result in error propagation among different steps because APC is affected by ATE error. These methods are impractical for real-world scenarios where the ATE task is the base task for APC, and its result impacts the accuracy of APC. Thus, in this study, we focused on a multi-task learning model for Arabic ATE and APC in which the model is jointly trained on two subtasks simultaneously in a single model. This paper integrates the multi-task model, namely Local Cotext Foucse-Aspect Term Extraction and Polarity classification (LCF-ATEPC) and Arabic Bidirectional Encoder Representation from Transformers (AraBERT) as a shred layer for Arabic contextual text representation. The LCF-ATEPC model is based on a multi-head selfattention and local context focus mechanism (LCF) to capture the interactive information between an aspect and its context. Moreover, data augmentation techniques are proposed based on state-of-the-art augmentation techniques (word embedding substitution with constraints and contextual embedding (AraBERT)) to increase the diversity of the training dataset. This paper examined the effect of data augmentation on the multi-task model for Arabic ABSA. Extensive experiments were conducted on the original and combined datasets (merging the original and augmented datasets). Experimental results demonstrate that the proposed Multi-task model outperformed existing APC techniques. Superior results were obtained by AraBERT and LCF-ATEPC with fusion layer (AR-LCF-ATEPC-Fusion) and the proposed data augmentation word embedding-based method (FastText) on the combined dataset.

show abstract

Section: Performance Of Proposed Multi-task Model On the Original Dat...mentioning

confidence: 85%

Section: Performance Of Proposed Multi-task Model On the Original Dat...mentioning

confidence: 99%

Multi-Task Learning Model with Data Augmentation for Arabic Aspect-Based Sentiment Analysis

Fadel¹,

Abulnaja²,

Saleh³

2023

Computers, Materials &Amp; Continua

View full text Add to dashboard Cite

show abstract

“…Fadel et al [43] proposed the BF-BiLSTM-CRF model based on BERT to handle the target extraction task. They combined contextualized string embedding with the BERT language model to improve word representation as embedding layer, and on top of it they stacked two Bilstm layers with crf layer as output layer.…”

Section: Related Workmentioning

confidence: 99%

Arabic aspect sentiment polarity classification using BERT

2022

View full text Add to dashboard Cite

Aspect-based sentiment analysis (ABSA) is a textual analysis methodology that defines the polarity of opinions on certain aspects related to specific targets. The majority of research on ABSA is in English, with a small amount of work available in Arabic. Most previous Arabic research has relied on deep learning models that depend primarily on context-independent word embeddings (e.g. word2vec), where each word has a fixed representation independent of its context. This article explores the modeling capabilities of contextual embeddings from pre-trained language models, such as BERT, and making use of sentence pair input on Arabic aspect sentiment polarity classification task. In particular, we develop a simple but effective BERT-based neural baseline to handle this task. Our BERT architecture with a simple linear classification layer surpassed the state-of-the-art works, according to the experimental results on three different Arabic datasets. Achieving an accuracy of 89.51% on the Arabic hotel reviews dataset, 73.23% on the Human annotated book reviews dataset, and 85.73% on the Arabic news dataset.

show abstract

“…Intelligent education is an educational model on the campus, which provides convenience for students' study and life and improves children's ability to explore and understand new things and unknown worlds [9][10]. Now many colleges and universities have begun to use this teaching model to carry out curriculum reform.…”

Section: Intelligent Educationmentioning

confidence: 99%

Appliance of Deep Learning in Intelligent Education

2022

View full text Add to dashboard Cite

Intelligent education is the main direction of China's future development. In the information age, the traditional teaching mode can no longer meet the higher requirements for information and knowledge quality in students' learning. Based on this, this paper proposes a method to help improve neural network training speed, reduce training time and improve memory efficiency. First, it introduces the relevant theories of intelligent pedagogy through literature analysis, and then designs a set of experimental scheme according to the actual situation, and verifies its effectiveness and accuracy. Finally, the feasibility of the design scheme is verified by the experimental results. The test results show that the data processing time of the model is relatively fast, the delay time is relatively low, and the compatibility rate of the model is up to more than 90%. This provides a reference for further exploration of AI education in the future, and also contributes to the future development of China.

show abstract

Arabic Aspect Extraction Based on Stacked Contextualized Embedding With Deep Learning

Cited by 22 publications

References 58 publications

Multi-Task Learning Model with Data Augmentation for Arabic Aspect-Based Sentiment Analysis

Multi-Task Learning Model with Data Augmentation for Arabic Aspect-Based Sentiment Analysis

Arabic aspect sentiment polarity classification using BERT

Appliance of Deep Learning in Intelligent Education

Contact Info

Product

Resources

About