Exploring the frontiers of deep learning and natural language processing: A comprehensive overview of key challenges and emerging trends

Khan, Wahab; Daud, Ali; Khan, Khairullah; Muhammad, Shakoor; Haq, Rafiul

doi:10.1016/j.nlp.2023.100026

Cited by 14 publications

(7 citation statements)

References 84 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Unlike simpler NLP models, conversational agents offer better performance on various tasks [45]. However, more complex models also require high computational costs and large amounts of data for optimization [46,47], which may limit their adaptability to the different linguistic and cultural needs of different regions [48]. It is important to note that high-income countries have led research in this field and have advanced technological resources for developing these AI models compared to other low-and middle-income countries [49,50].…”

Section: Comparison With Other Studiesmentioning

confidence: 99%

Self-administered interventions based on natural language processing models for reducing depressive and anxious symptoms: Systematic review and meta-analysis (Preprint)

Villarreal-Zegarra,

García-Serna,

Quispe-Callo

et al. 2024

Preprint

View full text Add to dashboard Cite

BACKGROUND The introduction of Natural Language Processing (NLP) technologies has significantly enhanced the potential of self-directed interventions for treating anxiety and depression by improving human-computer interactions. Despite these advancements, particularly in AI and Large Language Models (LLMs), robust evidence validating their effectiveness remains sparse. OBJECTIVE To determine whether interventions based on NLP models can reduce depressive and anxiety symptoms. METHODS Our study was a systematic review, and the protocol was registered in PROSPERO (CRD42023472120). The databases we used for the systematic review are Web of Science, SCOPUS, MEDLINE (via PubMed), PsycINFO (via EBSCO), IEEE Xplore, EMBASE (via EBSCO), and Cochrane Library. The quality of the included studies was assessed using the JBI Critical Appraisal Tools. RESULTS 21 articles were selected for review, and 16 were included in the meta-analysis for each outcome. The overall meta-analysis showed that self-administered interventions based on NLP models were significantly more effective in reducing depressive symptoms (SMD=0.819; 95%CI: 0.389-1.250; p<0.001) and anxiety symptoms (SMD=0.272; 95% CI: 0.116-0.428; p=0.001) compared with various control conditions. In subgroup analysis, AI-based NLP was shown to be effective in reducing depressive (SMD=1.059 [0.520 to 1.597]; p<0.001) and anxiety symptoms (SMD=0.302 [0.073 to 0.532]; p=0.010) compared with pooled control conditions. Also, NLP-based interventions overall outperform psychoeducation and bibliotherapy in reducing both depressive (SMD=1.481 [0.368 to 2.594]; p=0.009) and anxiety symptoms (SMD=0.561 [0.195 to 0.927]; p=0.003). In addition, these interventions are more effective than waitlist or no intervention in reducing anxious symptoms (SMD=0.196 [0.042 to 0.351]; p=0.013). CONCLUSIONS Our findings support the usefulness of self-applied NLP-based interventions in alleviating widely prevalent mental health problems such as depressive and anxious symptoms. CLINICALTRIAL Protocol was registered in PROSPERO (CRD42023472120)

show abstract

Section: Comparison With Other Studiesmentioning

confidence: 99%

Self-administered interventions based on natural language processing models for reducing depressive and anxious symptoms: Systematic review and meta-analysis (Preprint)

Villarreal-Zegarra,

García-Serna,

Quispe-Callo

et al. 2024

Preprint

View full text Add to dashboard Cite

show abstract

“…The effectiveness of domain-adaptive pre-training, which involves additional pre-training steps on domain-specific data, has been consistently highlighted as a strategy to bridge the gap between general corpora and specialized tasks [75], [76]. Additionally, the exploration of novel architectures and learning algorithms suggests ongoing innovation in transfer learning strategies, aiming to address its challenges and maximize its benefits [10], [71], [77].…”

Section: Transfer Learning Techniques and Strategiesmentioning

confidence: 99%

Evaluation of Transfer Learning and Adaptability in Large Language Models with the GLUE Benchmark

Sulaiman,

Hamzah

2024

Preprint

View full text Add to dashboard Cite

show abstract

“…These studies outperform the rule-based approach but the success of these models heavily depends on the selection of the features that are derived from the annotated corpus and are used in the training process. Deep learning (DL) based studies use different pre-trained word embedding techniques [21][22][23][24][25][26][27] to map the words in vectors using the language vocabulary to automatically extract meaningful relationships among words in the dataset. Due to limited vocabulary size, out-of-vocabulary words pose significant challenges for morphology-rich languages [28] like Urdu due to language complexities [29].…”

Section: Introductionmentioning

confidence: 99%

A deep learning approach for Named Entity Recognition in Urdu language

Anam,

Anwar,

Jamal

et al. 2024

PLoS ONE

View full text Add to dashboard Cite

Named Entity Recognition (NER) is a natural language processing task that has been widely explored for different languages in the recent decade but is still an under-researched area for the Urdu language due to its rich morphology and language complexities. Existing state-of-the-art studies on Urdu NER use various deep-learning approaches through automatic feature selection using word embeddings. This paper presents a deep learning approach for Urdu NER that harnesses FastText and Floret word embeddings to capture the contextual information of words by considering the surrounding context of words for improved feature extraction. The pre-trained FastText and Floret word embeddings are publicly available for Urdu language which are utilized to generate feature vectors of four benchmark Urdu language datasets. These features are then used as input to train various combinations of Long Short-Term Memory (LSTM), Bidirectional LSTM (BiLSTM), Gated Recurrent Unit (GRU), CRF, and deep learning models. The results show that our proposed approach significantly outperforms existing state-of-the-art studies on Urdu NER, achieving an F-score of up to 0.98 when using BiLSTM+GRU with Floret embeddings. Error analysis shows a low classification error rate ranging from 1.24% to 3.63% across various datasets showing the robustness of the proposed approach. The performance comparison shows that the proposed approach significantly outperforms similar existing studies.

show abstract

Exploring the frontiers of deep learning and natural language processing: A comprehensive overview of key challenges and emerging trends

Cited by 14 publications

References 84 publications

Self-administered interventions based on natural language processing models for reducing depressive and anxious symptoms: Systematic review and meta-analysis (Preprint)

Self-administered interventions based on natural language processing models for reducing depressive and anxious symptoms: Systematic review and meta-analysis (Preprint)

Evaluation of Transfer Learning and Adaptability in Large Language Models with the GLUE Benchmark

A deep learning approach for Named Entity Recognition in Urdu language

Contact Info

Product

Resources

About