Change search
Link to record
Permanent link

Direct link
Publications (10 of 18) Show all publications
Imbert, F., Corbillé, S., Han, H. & Barney Smith, E. H. (2026). Domain Adaptation Based Pipeline for Character Classification and Handwritten Text Recognition. International Journal on Document Analysis and Recognition
Open this publication in new window or tab >>Domain Adaptation Based Pipeline for Character Classification and Handwritten Text Recognition
2026 (English)In: International Journal on Document Analysis and Recognition, ISSN 1433-2833, E-ISSN 1433-2825Article in journal (Refereed) Epub ahead of print
Abstract [en]

Character classification and handwritten text recognition, particularly in specialized tasks, often suffer from limited labeled data, while a large amount of unlabeled data remains unused. To address this challenge, we propose a novel unsupervised domain adaptation-based pipeline to reduce the distribution gap between source and target domains. Our pipeline follows three main steps. In the first step, we pretrained an expert model on source data, to learn a strong general representation. In the second step, we perform unsupervised domain adaptation to leverage unlabeled target data, adapting the model to the specific characteristics of the target domain. In the third step, we trained the expert model on source data in the target style. By aligning the feature distributions between the source and target, our method enables the model to generalize effectively across diverse datasets. The unsupervised training increases the computation time only during training not in inference. We evaluate the pipeline on both character-level and historical handwritten line recognition tasks, demonstrating its flexibility. We establish a new state-of-the-art on the Safran-MNIST-DLS dataset defined in Track 2 of the DAGECC competition with 76.17% Macro F1-score and achieve competitive results on each of the specific datasets of READ 2018 without using the labels of the target data. Code is available at https://github.com/simon-corbi/DRANet-for-classification-and-htr

Place, publisher, year, edition, pages
Springer Nature, 2026
Keywords
Domain Adaptation, Character Classification, Handwritten Text Recognition, Safran-MNIST-DLS, READ 2018
National Category
Artificial Intelligence Natural Language Processing
Research subject
Machine Learning
Identifiers
urn:nbn:se:ltu:diva-116983 (URN)10.1007/s10032-026-00575-9 (DOI)001728417000001 ()2-s2.0-105034750858 (Scopus ID)
Funder
Knut and Alice Wallenberg FoundationThe Kempe Foundations, JCSMK23-0109
Note

Full text license: CC BY 4.0; 

Available from: 2026-04-07 Created: 2026-04-07 Last updated: 2026-07-02Bibliographically approved
Han, H., Trimi, S. & Lee, S. M. (2026). Tiny Machine Learning (TinyML): Research trends and future application opportunities. Array, 29, Article ID 100674.
Open this publication in new window or tab >>Tiny Machine Learning (TinyML): Research trends and future application opportunities
2026 (English)In: Array, E-ISSN 2590-0056, Vol. 29, article id 100674Article in journal (Refereed) Published
Abstract [en]

Tiny Machine Learning (TinyML) enables artificial intelligence on low-power edge devices, yet a quantitative understanding of TinyML research remains limited. This study addresses this gap through a comprehensive bibliometric analysis of 392 peer-reviewed publications (2020–2024) from the Web of Science, using Biblioshiny and VOSviewer. This article contributes by mapping the first bibliometric structure of TinyML, identifying major trends (exponential publication growth, strong international collaboration, core research themes, key contributors, etc.) and proposing future directions (such as sustainable hardware, federated learning, ethical frameworks, etc.). The findings provide a scholarly foundation and strategic roadmap for advancing scalable, energy-efficient, and privacy-preserving TinyML applications.

Place, publisher, year, edition, pages
Elsevier B.V., 2026
Keywords
Tiny machine learning, Bibliometric analysis, Web of science, Biblioshiny, VOSviewer, Edge computing
National Category
Computer Sciences
Research subject
Machine Learning
Identifiers
urn:nbn:se:ltu:diva-116160 (URN)10.1016/j.array.2025.100674 (DOI)001662316800001 ()2-s2.0-105026656074 (Scopus ID)
Funder
Knut and Alice Wallenberg Foundation
Note

Full text: CC BY license;

Available from: 2026-01-26 Created: 2026-01-26 Last updated: 2026-06-30Bibliographically approved
Imbert, F., Adewumi, O. & Han, H. (2025). A Novel Preprocessing-Driven Approach to Remaining Useful Life (RUL) Prediction Using Temporal Convolutional Networks (TCN). In: 2025 IEEE 37th International Conference on Tools with Artificial Intelligence (ICTAI): . Paper presented at 2025 IEEE 37th International Conference on Tools with Artificial Intelligence (ICTAI 2025), Athens, Greece, November 3-5, 2025 (pp. 1095-1102). IEEE
Open this publication in new window or tab >>A Novel Preprocessing-Driven Approach to Remaining Useful Life (RUL) Prediction Using Temporal Convolutional Networks (TCN)
2025 (English)In: 2025 IEEE 37th International Conference on Tools with Artificial Intelligence (ICTAI), IEEE, 2025, p. 1095-1102Conference paper, Published paper (Refereed)
Abstract [en]

Accurate prediction of Remaining Useful Life (RUL) in aero-engines is vital for predictive maintenance, improved operational reliability, and reduced lifecycle costs. While deep learning approaches have demonstrated strong potential in this area, most existing methods focus primarily on model architecture design and treat input features uniformly, often neglecting the influence of data preprocessing. In this work, we propose a novel preprocessing pipeline that enhances RUL prediction by improving data quality and temporal representation before model training. Our approach leverages complete temporal sequences and generates RUL estimates at each timestep, enabling the model to capture finegrained degradation dynamics and deliver continuous prognostic insights throughout the engine's operational life. To validate the effectiveness of the proposed pipeline, we conduct experiments on the NASA C-MAPSS dataset. Comparative evaluations against a suite of state-of-the-art neural models including CNN, RNN, LSTM, DCNN, TCN, BiGRUTSAM, AGCNN, and ATCN, demonstrate that our approach consistently achieves superior accuracy and robustness in aeroengine RUL prediction. These results highlight the critical role of preprocessing in maximizing the effectiveness of neural prognostic models.

Place, publisher, year, edition, pages
IEEE, 2025
Keywords
Remaining useful life prediction, Temporal convolutional network, data preprocessing
National Category
Reliability and Maintenance Artificial Intelligence
Research subject
Machine Learning
Identifiers
urn:nbn:se:ltu:diva-116187 (URN)10.1109/ictai66417.2025.00160 (DOI)001813729300149 ()2-s2.0-105031904850 (Scopus ID)
Conference
2025 IEEE 37th International Conference on Tools with Artificial Intelligence (ICTAI 2025), Athens, Greece, November 3-5, 2025
Funder
The Kempe Foundations, CSMK23-0109
Note

ISBN for host publication: 979-8-3315-4919-0

Available from: 2026-01-27 Created: 2026-01-27 Last updated: 2026-11-03Bibliographically approved
Han, H. & Li, J. (2025). Leveraging industry 4.0 technologies for sustainable supply chain management: A systematic review. Cleaner Logistics and Supply Chain, 16, Article ID 100246.
Open this publication in new window or tab >>Leveraging industry 4.0 technologies for sustainable supply chain management: A systematic review
2025 (English)In: Cleaner Logistics and Supply Chain, ISSN 2772-3909, Vol. 16, article id 100246Article, review/survey (Refereed) Published
Abstract [en]

Integrating sustainability into supply chain management is both a challenge and a strategic necessity. This review aims to bridge the gap in the literature by examining how Industry 4.0 (I4.0) technologies can advance sustainable supply chain management (SSCM). By systematically analyzing 231 studies, this paper identifies major technological barriers, including eco-efficiency deficits, data privacy concerns, and compatibility issues, that hinder SSCM adoption. It systematically explores six key I4.0 technologies—Blockchain, Big Data Analytics, IoT, Additive Manufacturing, Cyber-Physical Systems, and Cloud Computing—demonstrating their capabilities to overcome these challenges and achieve sustainable outcomes. Furthermore, this review highlights critical research gaps and proposes a detailed roadmap for future exploration, emphasizing the development of integration frameworks, advanced data management, and the exploration of emerging technologies like AI and augmented reality. By offering a strategic framework on I4.0′s potential to drive sustainable practices, this study provides actionable insights for scholars, practitioners, and policymakers to leverage I4.0 technologies for sustainable innovation.

Place, publisher, year, edition, pages
Elsevier BV, 2025
Keywords
Sustainable Supply Chain Management (SSCM), Industry 4.0 (I4.0), Systematic Review, Framework, Sustainability, Technologies
National Category
Production Engineering, Human Work Science and Ergonomics
Research subject
Machine Learning
Identifiers
urn:nbn:se:ltu:diva-114134 (URN)10.1016/j.clscn.2025.100246 (DOI)001554603200001 ()2-s2.0-105010861247 (Scopus ID)
Funder
Knut and Alice Wallenberg Foundation
Note

Validerad;2025;Nivå 1;2025-08-14 (u5);

Full text license: CC BY

Available from: 2025-08-01 Created: 2025-08-01 Last updated: 2025-11-28Bibliographically approved
Han, H. & Trimi, S. (2024). Analysis of cloud computing-based education platforms using unsupervised random forest. Education and Information Technologies: Official Journal of the IFIP technical committee on Education, 29, 15905-15932
Open this publication in new window or tab >>Analysis of cloud computing-based education platforms using unsupervised random forest
2024 (English)In: Education and Information Technologies: Official Journal of the IFIP technical committee on Education, ISSN 1360-2357, E-ISSN 1573-7608, Vol. 29, p. 15905-15932Article in journal (Refereed) Published
Abstract [en]

Cloud computing-based online education has played a vital role in enabling uninterrupted learning during crises such as the COVID-19 pandemic. This study explored the key variables associated with cloud computing that can effectively support the operation of online education platforms. By analyzing real data from 63 online learning platforms, the study utilizes the unsupervised random forest (RF) method, ranking the importance of cloud computing-related variables using the Gini index and permutation importance and Boruta. The study's contributions include the application of innovative techniques such as Addcl1 and K-means clustering for data pre-processing, unsupervised RF dissimilarity calculated by partitioning around medoids (PAM) clustering and multidimensional scaling (MDS) plots, and the comparison of RF accuracy (98.4%) with other machine learning techniques using WEKA software. The results highlight variables such as public cloud, commercial sourcing, course number limitations, and synchronized tools as important factors in the successful implementation of cloud computing-based online learning platforms. The theoretical and practical implications of the study results can be used by researchers and practitioners to structure cloud computing-based online learning platforms that could successfully function for distance learning during crises like the COVID-19 pandemic.

Place, publisher, year, edition, pages
Springer Nature, 2024
National Category
Computer Sciences
Identifiers
urn:nbn:se:ltu:diva-110029 (URN)10.1007/s10639-024-12457-w (DOI)001156356300005 ()2-s2.0-85184180369 (Scopus ID)
Available from: 2024-09-19 Created: 2024-09-19 Last updated: 2025-12-03Bibliographically approved
Han, H. & Trimi, S. (2022). Cloud Computing-based Higher Education Platforms during the COVID-19 Pandemic. In: IC4E '22: Proceedings of the 2022 13th International Conference on E-Education, E-Business, E-Management, and E-Learning: . Paper presented at 2022 13th International Conference on E-Education, E-Business, E-Management, and E-Learning (IC4E 2022), January 14-17, 2022, Tokyo, Japan (pp. 83-89). , 16
Open this publication in new window or tab >>Cloud Computing-based Higher Education Platforms during the COVID-19 Pandemic
2022 (English)In: IC4E '22: Proceedings of the 2022 13th International Conference on E-Education, E-Business, E-Management, and E-Learning, 2022, Vol. 16, p. 83-89Conference paper, Published paper (Refereed)
National Category
Computer Sciences
Identifiers
urn:nbn:se:ltu:diva-110032 (URN)10.1145/3514262.3514307 (DOI)001107411100013 ()2-s2.0-85129180904 (Scopus ID)
Conference
2022 13th International Conference on E-Education, E-Business, E-Management, and E-Learning (IC4E 2022), January 14-17, 2022, Tokyo, Japan
Available from: 2024-09-19 Created: 2024-09-19 Last updated: 2025-10-21Bibliographically approved
Han, H. & Jingyue, L. (2022). Tiny Machine Learning: A New Technique for AI Security. ERCIM News, 129, 16-17
Open this publication in new window or tab >>Tiny Machine Learning: A New Technique for AI Security
2022 (English)In: ERCIM News, ISSN 0926-4981, E-ISSN 1564-0094, Vol. 129, p. 16-17Article in journal (Refereed) Published
Place, publisher, year, edition, pages
ECRIM, 2022
National Category
Computer Sciences
Identifiers
urn:nbn:se:ltu:diva-110031 (URN)
Available from: 2024-09-19 Created: 2024-09-19 Last updated: 2025-10-21Bibliographically approved
Han, H. & Siebert, J. (2022). TinyML: A Systematic Review and Synthesis of Existing Research. In: 2022 International Conference on Artificial Intelligence in Information and Communication (ICAIIC): . Paper presented at The 4th International Conference on Artificial Intelligence in Information and Communication, February 21-24, 2022, Shilla Stay Jeju, Jeju lsland, Korea (pp. 269-274).
Open this publication in new window or tab >>TinyML: A Systematic Review and Synthesis of Existing Research
2022 (English)In: 2022 International Conference on Artificial Intelligence in Information and Communication (ICAIIC), 2022, p. 269-274Conference paper, Published paper (Refereed)
National Category
Computer Sciences
Identifiers
urn:nbn:se:ltu:diva-110034 (URN)10.1109/icaiic54071.2022.9722636 (DOI)2-s2.0-85127720179 (Scopus ID)
Conference
The 4th International Conference on Artificial Intelligence in Information and Communication, February 21-24, 2022, Shilla Stay Jeju, Jeju lsland, Korea
Available from: 2024-09-19 Created: 2024-09-19 Last updated: 2025-10-21Bibliographically approved
Han, H. & Trimi, S. (2022). Towards a data science platform for improving SME collaboration through Industry 4.0 technologies. Technological forecasting & social change, 174, Article ID 121242.
Open this publication in new window or tab >>Towards a data science platform for improving SME collaboration through Industry 4.0 technologies
2022 (English)In: Technological forecasting & social change, ISSN 0040-1625, E-ISSN 1873-5509, Vol. 174, article id 121242Article in journal (Refereed) Published
Place, publisher, year, edition, pages
Elsevier, 2022
National Category
Computer Sciences
Identifiers
urn:nbn:se:ltu:diva-110030 (URN)10.1016/j.techfore.2021.121242 (DOI)000709726300010 ()2-s2.0-85115968241 (Scopus ID)
Available from: 2024-09-19 Created: 2024-09-19 Last updated: 2025-12-03Bibliographically approved
Han, H. (2021). Standardisation for Security Applications and Technologies. ERCIM News, 126, 42-43
Open this publication in new window or tab >>Standardisation for Security Applications and Technologies
2021 (English)In: ERCIM News, ISSN 0926-4981, E-ISSN 1564-0094, Vol. 126, p. 42-43Article in journal (Refereed) Published
National Category
Computer Sciences
Identifiers
urn:nbn:se:ltu:diva-110069 (URN)
Available from: 2024-09-23 Created: 2024-09-23 Last updated: 2025-10-21Bibliographically approved
Organisations
Identifiers
ORCID iD: ORCID iD iconorcid.org/0000-0001-9320-3808

Search in DiVA

Show all publications