Skip to content

SS08

SS08: Innovations in Embedded Lightweight Machine Learning Models: Enabling TinyML at the Edge for Practical Applications

Absalom El-Shamir Ezugwu

North-West University, South Africa

Rytis Paškauskas

The Abdus Salam International Centre for Theoretical Physics (ICTP), Trieste, Italy

Diego Oliva

Universidad de Guadalajara, Mexico

Partha Pratim Ray

Sikkim University, India

Wen-Sheng Zhao

Hangzhou Dianzi University, China

Seyed Jalaleddin Mousavirad

Mid Sweden University, Sweden

Abstract

The rapid growth of machine learning and artificial intelligence has led to groundbreaking solutions across various domains. However, deploying these models on resource-constrained edge devices such as microcontrollers, sensors, and embedded systems remains a significant challenge. TinyML, a field at the intersection of machine learning and embedded systems, focuses on developing lightweight, efficient, and optimized ML models for such devices, enabling real-time decision-making while conserving power and memory resources.

This workshop will explore practical advancements, techniques, and applications of TinyML in addressing real-world challenges. Topics will cover the full spectrum of TinyML development, from model optimization techniques and hardware advancements to impactful case studies in diverse fields like smart agriculture, healthcare, environmental monitoring, and industrial IoT.

The special session aims to create a valuable platform for researchers and industry partners to present and discuss cutting-edge advancements, emerging trends, and innovative solutions in the field of TinyML.

This special session will include (but is not limited to) the following topics:

  • Model Optimization for TinyML
    • Quantization, pruning, and compression of ML models for edge deployment.
    • Lightweight neural network design and Neural Architecture Search (NAS).
    • Transfer learning and model distillation for TinyML.
  • Hardware and Software Integration
    • Advances in hardware platforms (microcontrollers, accelerators, etc.) for TinyML.
    • Software frameworks and tools: TensorFlow Lite Micro, Edge Impulse, and PyTorch Mobile.
    • Benchmarking TinyML performance on resource-constrained devices.
  • Energy Efficiency and Sustainability
    • Low-power AI techniques for energy-constrained devices.
    • Renewable energy integration into TinyML-powered IoT systems.
    • Power-aware ML model optimization.
  • Real-World Applications of TinyML
    • Smart Agriculture: Precision farming, pest detection, and soil health monitoring.
    • Healthcare: Remote patient monitoring, wearable medical devices, and diagnostics.
    • Environmental Monitoring: Edge-based climate change sensors, pollution detection, and wildlife tracking.
    • Industrial IoT: Predictive maintenance, fault detection, and process automation.

Organizers

Prof. Absalom El-Shamir Ezugwu earned his B.Sc. in mathematics with computer science, followed by M.Sc. and Ph.D. degrees in computer science from Ahmadu Bello University Zaria, Nigeria. He is a full Professor of Computer Science within the Unit for Data Science and Computing at North-West University Potchefstroom, South Africa. Absalom has successfully mentored and graduated numerous PhD and MSc students in computer science, achieving a notable record of high-impact publications in respected journals and conferences. He is also a visiting research Professor with the School of Computer Science and Applied Mathematics, University of the Witwatersrand, Johannesburg, South Africa. His research focuses on artificial intelligence, machine learning, deep learning, evolutionary computation, swarm intelligence, and nature-inspired algorithm design, with a specific emphasis on computational intelligence and metaheuristic solutions for real-world global optimization problems.

Dr. Rytis Paškauskas is a researcher at the Abdus Salam International Centre for Theoretical Physics (ICTP) in Trieste, Italy, where he works on embedded machine-learning projects and IoT applications. He received his PhD degree in physics from Georgia Institute of Technology (USA) and has since worked at internationally renowned institutions, such as ELETTRA Sincrotrone (Italy), CNR Pisa (Italy), CNRS Lyon (France), and the National Institute for Theoretical Physics in Stellenbosch (South Africa).

Dr. Diego Oliva received a B.S. degree in Electronics and Computer Engineering from the Industrial Technical Education Center (CETI) of Guadalajara, Mexico, in 2007 and an M.Sc. degree in Electronic Engineering and Computer Sciences from the University of Guadalajara, Mexico, in 2010. He obtained a Ph. D. in Informatics in 2015 from the Universidad Complutense de Madrid. Currently, he is an Associate Professor at the University of Guadalajara in Mexico. He is a member of the Mexican National Research System (SNII), a Senior member of the IEEE, and a member of the Mexican Academy of Computer Sciences (AMEXCOMP). He is an editor of multiple journals including Swarm and Evolutionary Computation; he also actively participates and organizes special tracks in conferences such as Evostar, CEC, and WCCI. His research interests include evolutionary and swarm algorithms, hybridization of evolutionary and swarm algorithms, and computational intelligence.

Dr. Partha Pratim Ray is an accomplished academic and researcher with over 12 years of teaching and research experience in the Department of Computer Applications at Sikkim University, India. Recognized among the World’s Top 2% Scientists by Stanford University multiple times, Dr. Ray has made significant contributions to the fields of the Internet of Things (IoT), Edge Computing, Pervasive Biomedical Informatics, and Pervasive Generative AI. He holds an M.Tech in Embedded Systems and is pursuing a PhD in “Enabling Large Language Models on Resource-Constrained Edge” at Sikkim University. Dr. Ray has published 108 SCI-indexed journal articles and multiple Scopus-indexed works, alongside several books and book chapters with leading publishers such as Elsevier and Springer. He has been elevated to Fellow, IETE, and is a Senior Member of IEEE and an active member of ACM. His research projects focus on deploying IoT-based solutions and lightweight communication frameworks for resource-constrained environments. Dr. Ray has also guided numerous master’s theses and mentored projects in cutting-edge areas like Blockchain, IoT-based healthcare systems, and edge computing. Committed to advancing technology and fostering innovation, Dr. Ray has received multiple awards, including the IEI Young Engineers Award and Emerald Literati Awards.

Prof. Wen-Sheng Zhao received a B.E. degree from the Harbin Institute of Technology, Harbin, China, in 2008, and a Ph.D. degree from Zhejiang University, Hangzhou, China, in 2013. He is currently a full professor at Hangzhou Dianzi University, Hangzhou. He has published three books, five chapters, and more than 150 SCI papers including more than 90 IEEE papers. His current research interests include modeling and simulation of integrated microsystems, and design of electromagnetic devices. Dr. Zhao is a senior member of IEEE and Chinese Institute of Electronics and serves as associate editor/editorial member/guest editor for several journals including Microelectronics Journal, IEEE Access, Chinese Journal of Electronics, and Micromachines.

Dr. Seyed Jalaleddin Mousavirad is currently a Postdoctoral researcher at Mid Sweden University, located in Sundsvall, Sweden. Previously, he served as a Research Fellow at the University of Beira Interior in Portugal, where he was actively involved in the European project called GreenStamp. Jalal obtained his PhD in Computer Engineering, specializing in Artificial Intelligence, from the University of Kashan in Iran. Following his doctoral studies, he worked at the University of Tehran (2018-2019) and Azad University (2019-2020) as an instructor. Additionally, he served as an Assistant Professor at the Faculty of Engineering at Hakim Sabzevari University in Iran. With a strong research background, Jalal has made significant contributions in the areas of pattern recognition, machine learning, image processing, and evolutionary computation. He has published six book chapters and over 100 papers in reputable academic journals and conferences. Jalal’s international research experiences also include visiting a world-class research group at Xi’an Jiaotong-Liverpool University in China He has organized several special sessions at prestigious conferences such as the IEEE CEC and EvoApplications.