A deep learning approach for Named Entity Recognition in Urdu language

Article Subjects > Engineering Europe University of Atlantic > Research > Scientific Production
Ibero-american International University > Research > Scientific Production
Ibero-american International University > Research > Articles and Books
Universidad Internacional do Cuanza > Research > Scientific Production
Fundación Universitaria Internacional de Colombia > Research > Scientific Production Open English 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. metadata Khan, Hikmat Ullah; Anam, Rimsha; Anwar, Muhammad Waqas; Jamal, Muhammad Hasan; Bajwa, Usama Ijaz; Diez, Isabel de la Torre; Silva Alvarado, Eduardo René; Soriano Flores, Emmanuel and Ashraf, Imran mail UNSPECIFIED, UNSPECIFIED, UNSPECIFIED, UNSPECIFIED, UNSPECIFIED, UNSPECIFIED, eduardo.silva@funiber.org, emmanuel.soriano@uneatlantico.es, UNSPECIFIED (2024) A deep learning approach for Named Entity Recognition in Urdu language. PLOS ONE, 19 (3). e0300725. ISSN 1932-6203

Text
journal.pone.0300725.pdf
Available under License Creative Commons Attribution.
Download (1MB)

Official URL: http://doi.org/10.1371/journal.pone.0300725

Abstract

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.

Document Type:	Article
Subject classification:	Subjects > Engineering
Divisions:	Europe University of Atlantic > Research > Scientific Production Ibero-american International University > Research > Scientific Production Ibero-american International University > Research > Articles and Books Universidad Internacional do Cuanza > Research > Scientific Production Fundación Universitaria Internacional de Colombia > Research > Scientific Production
Deposited:	28 May 2024 23:30
Last Modified:	09 Dec 2024 23:30
URI:	https://repositorio.unib.org/id/eprint/12369

Actions (login required)

View Object

open

Benchmarking multiple instance learning architectures from patches to pathology for prostate cancer detection and grading using attention-based weak supervision

Histopathological evaluation is necessary for the diagnosis and grading of prostate cancer, which is still one of the most common cancers in men globally. Traditional evaluation is time-consuming, prone to inter-observer variability, and challenging to scale. The clinical usefulness of current AI systems is limited by the need for comprehensive pixel-level annotations. The objective of this research is to develop and evaluate a large-scale benchmarking study on a weakly supervised deep learning framework that minimizes the need for annotation and ensures interpretability for automated prostate cancer diagnosis and International Society of Urological Pathology (ISUP) grading using whole slide images (WSIs). This study rigorously tested six cutting-edge multiple instance learning (MIL) architectures (CLAM-MB, CLAM-SB, ILRA-MIL, AC-MIL, AMD-MIL, WiKG-MIL), three feature encoders (ResNet50, CTransPath, UNI2), and four patch extraction techniques (varying sizes and overlap) using the PANDA dataset (10,616 WSIs), yielding 72 experimental configurations. The methodology used distributed cloud computing to process over 31 million tissue patches, implementing advanced attention mechanisms to ensure clinical interpretability through Grad-CAM visualizations. The optimum configuration (UNI2 encoder with ILRA-MIL, 256 256 patches, 50% overlap) achieved 78.75% accuracy and 90.12% quadratic weighted kappa (QWK), outperforming traditional methods and approaching expert pathologist-level diagnostic capability. Overlapping smaller patches offered the best balance of spatial resolution and contextual information, while domain-specific foundation models performed noticeably better than generic encoders. This work is the first large-scale, comprehensive comparison of weekly supervised MIL methods for prostate cancer diagnosis and grading. The proposed approach has excellent clinical diagnostic performance, scalability, practical feasibility through cloud computing, and interpretability using visualization tools.

Producción Científica

Naveed Anwer Butt mail , Dilawaiz Sarwat mail , Irene Delgado Noya mail irene.delgado@uneatlantico.es, Kilian Tutusaus mail kilian.tutusaus@uneatlantico.es, Nagwan Abdel Samee mail , Imran Ashraf mail ,

Butt

open

A scalable and secure federated learning authentication scheme for IoT

Secure and scalable authentication remains a fundamental challenge in Internet of Things (IoT) networks due to constrained device resources, dynamic topology, and the absence of centralized trust infrastructures. Conventional password-based and certificate-driven authentication schemes incur high computation, storage, and communication overhead, limiting their suitability for large-scale deployments. To address these limitations, this paper proposes ScLBS, a federated learning (FL)–based self-certified authentication scheme for distributed and sustainable IoT environments. ScLBS integrates self-certified public key cryptography with FL-driven trust adaptation, enabling decentralized public key derivation without reliance on third-party certificate authorities or exposure of private credentials. A zero-knowledge mechanism combined with location-aware authentication strengthens resistance to impersonation, Sybil, and replay attacks. Hierarchical key management supported by a -tree enables efficient group rekeying and preserves forward and backward secrecy under dynamic membership. Formal security verification is conducted under the Dolev–Yao adversary model using ProVerif, confirming secrecy of private and session keys (SKs) and correctness of authentication. Extensive NS-3 simulations and ablation analysis demonstrate that ScLBS achieves lower authentication delay, reduced message overhead, improved network utilization, and decreased energy consumption compared to representative IoT authentication schemes, while maintaining bounded FL overhead. These results indicate that ScLBS provides a balanced trade-off between security strength, scalability, and resource efficiency for constrained IoT networks.

Producción Científica

Premkumar Chithaluru mail , B. Veera Jyothi mail , Fahd S. Alharithi mail , Wojciech Ksiazek mail , M. Ramchander mail , Aman Singh mail aman.singh@uneatlantico.es, Ravi Kumar Rachavaram mail ,

Chithaluru

open

Innovative Application of Chatbots in Clinical Nutrition Education: The E+DIEting_Lab Experience in University Students

Background/Objectives: The growing integration of Artificial Intelligence (AI) and chatbots in health professional education offers innovative methods to enhance learning and clinical preparedness. This study aimed to evaluate the educational impact and perceptions in university students of Human Nutrition and Dietetics, regarding the utility, usability, and design of the E+DIEting_Lab chatbot platform when implemented in clinical nutrition training. Methods: The platform was piloted from December 2023 to April 2025 involving 475 students from multiple European universities. While all 475 students completed the initial survey, 305 finished the follow-up evaluation, representing a 36% attrition rate. Participants completed surveys before and after interacting with the chatbots, assessing prior experience, knowledge, skills, and attitudes. Data were analyzed using descriptive statistics and independent samples t-tests to compare pre- and post-intervention perceptions. Results: A total of 475 university students completed the initial survey and 305 the final evaluation. Most university students were females (75.4%), with representation from six languages and diverse institutions. Students reported clear perceived learning gains: 79.7% reported updated practical skills in clinical dietetics and communication were updated, 90% felt that new digital tools improved classroom practice, and 73.9% reported enhanced interpersonal skills. Self-rated competence in using chatbots as learning tools increased significantly, with mean knowledge scores rising from 2.32 to 2.66 and skills from 2.39 to 2.79 on a 0–5 Likert scale (p < 0.001 for both). Perceived effectiveness and usefulness of chatbots as self-learning tools remained positive but showed a small decline after use (effectiveness from 3.63 to 3.42; usefulness from 3.63 to 3.45), suggesting that hands-on experience refined, but did not diminish, students’ overall favorable views of the platform. Conclusions: The implementation and pilot evaluation of the E+DIEting_Lab self-learning virtual patient chatbot platform demonstrate that structured digital simulation tools can significantly improve perceived clinical nutrition competences. These findings support chatbot adoption in dietetics curricula and inform future digital education innovations.

Producción Científica

Iñaki Elío Pascual mail inaki.elio@uneatlantico.es, Kilian Tutusaus mail kilian.tutusaus@uneatlantico.es, Imanol Eguren García mail imanol.eguren@uneatlantico.es, Álvaro Lasarte García mail , Arturo Ortega-Mansilla mail arturo.ortega@uneatlantico.es, Thomas Prola mail thomas.prola@uneatlantico.es, Sandra Sumalla Cano mail sandra.sumalla@uneatlantico.es,

Elío Pascual

open

Suicide Ideation Detection Using Social Media Data and Ensemble Machine Learning Model

Identifying the emotional state of individuals has useful applications, particularly to reduce the risk of suicide. Users’ thoughts on social media platforms can be used to find cues on the emotional state of individuals. Clinical approaches to suicide ideation detection primarily rely on evaluation by psychologists, medical experts, etc., which is time-consuming and requires medical expertise. Machine learning approaches have shown potential in automating suicide detection. In this regard, this study presents a soft voting ensemble model (SVEM) by leveraging random forest, logistic regression, and stochastic gradient descent classifiers using soft voting. In addition, for the robust training of SVEM, a hybrid feature engineering approach is proposed that combines term frequency-inverse document frequency and the bag of words. For experimental evaluation, “Suicide Watch” and “Depression” subreddits on the Reddit platform are used. Results indicate that the proposed SVEM model achieves an accuracy of 94%, better than existing approaches. The model also shows robust performance concerning precision, recall, and F1, each with a 0.93 score. ERT and deep learning models are also used, and performance comparison with these models indicates better performance of the SVEM model. Gated recurrent unit, long short-term memory, and recurrent neural network have an accuracy of 92% while the convolutional neural network obtains an accuracy of 91%. SVEM’s computational complexity is also low compared to deep learning models. Further, this study highlights the importance of explainability in healthcare applications such as suicidal ideation detection, where the use of LIME provides valuable insights into the contribution of different features. In addition, k-fold cross-validation further validates the performance of the proposed approach.

Producción Científica

Erol KINA mail , Jin-Ghoo Choi mail , Abid Ishaq mail , Rahman Shafique mail , Mónica Gracia Villar mail monica.gracia@uneatlantico.es, Eduardo René Silva Alvarado mail eduardo.silva@funiber.org, Isabel de la Torre Diez mail , Imran Ashraf mail ,

KINA

open

In silico prediction, molecular docking and simulation of natural flavonoid apigenin and xanthoangelol E against human metapneumovirus

Human metapneumovirus (hMPV) is one of the potential pandemic pathogens, and it is a concern for elderly subjects and immunocompromised patients. There is no vaccine or specific antiviral available for hMPV. We conducted an in-silico study to predict initial antiviral candidates against human metapneumovirus. Our methodology included protein modeling, stability assessment, molecular docking, molecular simulation, analysis of non-covalent interactions, bioavailability, carcinogenicity, and pharmacokinetic profiling. We pinpointed four plant-derived bio-compounds as antiviral candidates. Among the compounds, apigenin showed the highest binding affinity, with values of − 8.0 kcal/mol for the hMPV-F protein and − 7.6 kcal/mol for the hMPV-N protein. Molecular dynamic simulations and further analyses confirmed that the protein-ligand docked complexes exhibited acceptable stability compared to two standard antiviral drugs. Additionally, these four compounds yielded satisfactory outcomes in bioavailability, drug-likeness, and ADME-Tox (absorption, distribution, metabolism, excretion, and toxicity) and STopTox analyses. This study highlights the potential of apigenin and xanthoangelol E as an initial antiviral candidate, underscoring the necessity for wet-lab evaluation, preclinical and clinical trials against human metapneumovirus infection.

Producción Científica

Hasan Huzayfa Rahaman mail , Afsana Khan mail , Nadim Sharif mail , Wasifuddin Ahmed mail , Nazmul Sharif mail , Rista Majumder mail , Silvia Aparicio Obregón mail silvia.aparicio@uneatlantico.es, Rubén Calderón Iglesias mail ruben.calderon@uneatlantico.es, Isabel De la Torre Díez mail , Shuvra Kanti Dey mail ,

Rahaman

Links of Interest

Links of Interest

A deep learning approach for Named Entity Recognition in Urdu language

Abstract

Actions (login required)

SUBJECT

ACCESS

LANGUAGE

Filters