Keynotes

Keynote Speakers

 

Prof. Vishal Monga, The Pennsylvania State University, USA
IEEE Fellow, AAIA Fellow

Biography: Dr. Vishal Monga is a faculty member in Electrical Engineering and Computer Science (EECS) at Pennsylvania State University, where he has been based since 2009. Before joining Penn State, he served as an Imaging Scientist at Xerox Research Labs and has held visiting positions at Microsoft Research and the University of Rochester. He holds a Ph.D. from the University of Texas, Austin, and completed his undergraduate studies at IIT Guwahati.

Prof. Monga’s research has earned him several distinguished honors, including the NSF CAREER Award, the Penn State Engineering Alumni Society Outstanding Research Award (2019), the PSEAS Premier Research Award (2022), and the IIT Guwahati Distinguished Alumnus Award (2022). He also received the Joel and Ruth Spira Teaching Excellence Award in 2016.

An active IEEE leader, Prof. Monga serves on multiple technical committees and has held editorial roles at leading journals including IEEE TIP, JSTSP, TCSVT, and SPL. He is a founding editorial member of Frontiers in Signal Processing and the editor of the Springer volume Handbook of Convex Optimization Methods in Imaging Science. Inducted into the National Academy of Inventors in 2022, he holds 45 U.S. patents. He is a Fellow of IEEE and of the Asia Pacific Artificial Intelligence Association (AAIA).

His research interests span Data Science, Artificial Intelligence, and Signal and Image Processing, with a focus on convex optimization methods in imaging and signal/image classification.

 

 

Prof. Liang Zhao, University of São Paulo, Brazil
the Stanford/Elsevier Top 2% Scientists List 2024

Biography: Liang Zhao is a Full Professor in the Department of Computing and Mathematics at University of São Paulo. He was included in the Stanford/Elsevier Top 2% Scientists List 2024 in the career-long category for Artificial Intelligence & Image Processing. He previously served as an Associate Editor for IEEE Transactions on Neural Networks and Learning Systems and currently serves as an Associate Editor for Neural Networks and IEEE Transactions on Systems, Man, and Cybernetics: Systems. His main research interests include complex network theory and machine learning, particularly deep learning, complex network-based machine learning, GNNs, LLMs, AI agents, and AI applications in agriculture. He has published more than 100 journal articles, many of them in influential journals such as Nature Communications, IEEE TNNLS, IEEE TEC, Neural Networks, and IEEE TKDE. In addition, he authored the book Machine Learning in Complex Networks, which became one of Springer's top 25% most downloaded books in 2017. He has also contributed to the editing of 9 books and published more than 160 conference papers.

 

 

Prof. Mauricio Ayala Rincón, Brasilia University, Brazil

 

 

 

 

 

Invited Speakers

Prof. Daniel Molina-Pérez, ESCOM, National Polytechnic Institute, Mexico

Speech Title: Noise: The Nightmare of Contrast Enhancement — Classical Techniques, Nature-Inspired Optimization, and Learning-Based Approaches

Abstract: Enhancing medical images is not only about improving contrast, but also about dealing with noise. In practice, increasing contrast often comes at the cost of amplifying noise, which can directly affect diagnostic reliability. Early methods improve contrast without explicitly accounting for this issue, while sequential approaches attempt to reduce noise and enhance contrast in separate steps. More advanced techniques introduce local, noise-aware adjustments during the enhancement process, achieving better results, although still relying on parameter tuning and limited transformation models.
Optimization-based approaches, particularly those inspired by nature, provide a way to explicitly control this trade-off, while deep learning methods learn it from data, often achieving strong performance but raising concerns such as limited controllability, hallucinations, and bias. Despite significant progress, balancing contrast and noise in a reliable and interpretable way remains a challenging problem.

Biography: DANIEL MOLINA-PÉREZ received the Ph.D. degree in Robotic and Mechatronic Systems from the Centro de Innovación y Desarrollo Tecnológico en Cómputo (CIDETEC), Instituto Politécnico Nacional, Mexico City, in 2024. He is currently a Professor-Researcher with the Escuela Superior de Cómputo (ESCOM), Instituto Politécnico Nacional. He is a member of the Artificial Intelligence and Data Science Network of the Instituto Politécnico Nacional. His research interests include evolutionary computation, multi-objective optimization, mixed-integer nonlinear programming, image enhancement, and hybrid optimization methods for solving complex engineering problems. Dr. Molina-Pérez has authored and co-authored multiple articles in indexed journals, including Swarm and Evolutionary Computation, PeerJ Computer Science, and Mathematical and Computational Applications, and has presented his work at leading international conferences such as the IEEE Congress on Evolutionary Computation (CEC). He received the Best Paper Award at the Mexican International Conference on Artificial Intelligence in 2022 and was recognized with the Outstanding Academic Performance Award at the doctoral level, reflecting his academic excellence and research contributions.

 

Prof. Anuranjan Misra, Noida International University, India

Speech Title: AI-Driven Digital Twin for Predictive Semiconductor Design using Hardware-in-the-Loop Learning

Abstract: The semiconductor systems we use today are very complicated which makes it really hard to predict how power they will use how well they will work and how long they will last when we are still designing them. The old tools we use to design these systems rely much on simple models that do not take into account how all the different parts of the system interact with each other and how they will work in the real world.
This is about a way of designing semiconductor systems using a digital twin that is controlled by artificial intelligence. This digital twin uses a kind of computer chip called an FPGA to test the system in real time and get data on how much power it uses how well it works and how the different parts of the system are working together. We use this data to train machine learning models that can learn how the design of the system affects how it works so we can predict what the system will be like before we even build it.
The digital twin is always getting better because it uses a loop where the artificial intelligence makes predictions we test them against data from the hardware and then the artificial intelligence makes new predictions based on what it has learned. This system uses a few kinds of machine learning including supervised learning to make predictions Bayesian methods to figure out how sure we are of our predictions and reinforcement learning to make the design better. By including rules about how the physical world works the system makes sure its predictions are realistic and stable.

Biography: Dr Anuranjan Misra is Professor & Dean at Greater Noida Institute of Technology(GNIOT), Greater Noida. He holds a Ph.D. in Computer Science & Engineering and possesses three Master's degrees in Computer Science and Engineering, Law, and Mathematics, demonstrating his multidisciplinary approach to education and research. He is EX -Chairman of Computer society of India, Ghaziabad Chapter. He is Head MSME Business Incubation-GNIOT Centre(an Initiative of Ministry of MSME, Govt. of India), Head MSME Design Centre-GNIOT Center(an Initiative of Ministry of MSME, Govt. of India), .President Institution's Innovation Council -GNIOT Centre(an Initiative of Ministry of Education, Govt. of India), Chair Unnat Bharat Abhiyan- GNIOT (an Initiative of Ministry of Education, Govt. of India), and Chair Smart Campus Cloud Network- GNIOT (an Initiative of TERRI & AICTE , New Delhi) . He has 25 years of rich experience in academics, research and industry. He has delivered more than 25 expert talks around the world. He has more than 150 publications. He has handled research funding of 6+ crores. He has Senior Member of ACM, IEEE, CSI, IACSIT, IACNG, IRACST, CSTA, ISOC, ICE, AEE, IFETS, ISMCDM, SIGSE. His research is in AI, Big Data, Cloud Computing, Data Science, and Algorithms. He is passionate about quality of higher education in India.

 

Assoc. Prof. Suraya Masrom, Universiti Teknologi Mara, Malaysia

Speech Title: Genetic Programming Based Automated Machine Learning

Abstract: Automated Machine Learning (AutoML) has emerged as a promising approach for solving classification and prediction problems, and continues to attract significant attention for further enhancement. One notable advancement in AutoML is the integration of evolutionary algorithms, particularly Genetic Programming (GP). GP facilitates the optimisation of machine learning pipelines by exploring and selecting the best combinations of algorithms and their corresponding hyperparameters.
As an evolutionary-based technique, the performance of GP in generating optimal pipelines is highly dependent on its parameterisation, especially mutation and crossover rates. This study investigates the effects of different combinations of mutation and crossover rates on AutoML performance across diverse datasets. The findings provide empirical insights supporting the notion that higher crossover rates tend to improve model accuracy, while lower crossover rates may lead to premature convergence.
In addition, this work introduces a novel software tool designed for the rapid implementation of GP-based AutoML, aimed at enhancing accessibility for non-expert users. This tool facilitates efficient experimentation and broadens the applicability of AutoML in various domains.

Biography: Suraya Masrom holds a Bachelor's Degree in Computer Science (Software Engineering) from Universiti Teknologi Malaysia and a Master's Degree in Computer Science from Universiti Putra Malaysia. She began her professional career in the industry as an Associate Network Engineer at Ramgate Systems Sdn. Bhd. in 1996. Following her postgraduate studies, she transitioned into academia as a full-time lecturer at Universiti Teknologi Malaysia, where she successfully completed two university-funded research projects within her first three years of service. In 2004, she joined Universiti Teknologi MARA (UiTM), Seri Iskandar, Perak, where she continues to serve as an academic in the field of Computer Science. Throughout her tenure at UiTM, she has led and contributed to more than 20 research projects funded by internal, national, international, and industry grants with a total value approaching RM1 million. Dr. Suraya is also actively engaged as a consultant in software systems and has contributed her expertise to various agencies, particularly in smart city planning and development initiatives. She has authored and co-authored over 100 publications in indexed journals, demonstrating her strong research impact. In addition to her research and consultancy roles, she has served as the Chief Editor of the Mathematical Sciences and Informatics Journal, and was honoured with the Best UiTM Journal Editor Award in 2021 in recognition of her editorial excellence.rmatic Journal and has been awarded as the the best UiTM Journal Editor in 2021.

Assoc. Prof. Jennifer O. Contreras, De La Salle University, Philippines

Speech Title: Ontology Learning in Disaster Risk Management Using a Hybrid SVM–Transformer Deep Learning Kernel

Abstract: Support Vector Machine (SVM) is a supervised learning algorithm widely used for classification tasks and capable of handling complex data through kernel functions. In ontology learning, SVM can classify concepts and relationships; however, traditional SVM kernels often struggle to capture deep contextual and semantic meanings in Disaster Risk Management (DRM) texts. This limitation may lead to inconsistent knowledge representation, reduced situational awareness, and suboptimal decision-making during mitigation, preparedness, response, and recovery phases.

To address this challenge, the study developed a hybrid SVM–Transformer Deep Learning (SVM-TDL) model that integrates neural projection learning and transformer-based contextual embeddings into the SVM framework. A curated DRM dataset was constructed from policy documents, hazard reports, research articles, and operational manuals. Domain-adapted BERT embeddings were generated and refined through a neural projection layer to form a transformer-informed neural kernel for SVM classification. The proposed hybrid model was evaluated against a classical SVM and a linear baseline using accuracy, F1-score, confusion matrices, kernel heatmaps, and conceptual cluster visualizations.

Experimental results demonstrated that the hybrid SVM-TDL model achieved 86.00% accuracy in ontology concept classification and 98.83% accuracy in relationship classification, significantly outperforming classical SVM (84.7% accuracy, 84.2% F1-score) and a linear baseline (78.9% accuracy). Misclassification among semantically similar DRM concepts—such as hazard, impact, exposure, and event was reduced by up to 35%, indicating improved non-linear separability and richer semantic discrimination. Kernel heatmaps and conceptual cluster visualizations further confirmed clearer and more stable boundary formation across concept classes. In addition, the hybrid model achieved improved computational efficiency, with time and space complexity O(n·d), substantially lower than the O(n²·d) and O(n²·d·p) complexities of RBF and polynomial kernels, enabling better scalability for large datasets. These findings indicate that integrating neural projection learning with Transformer-based representations substantially enhances SVM performance in DRM ontology learning. The resulting improvements contribute to more coherent ontologies, more transparent inter-agency communication, and more accurate decision-making across preparedness, response, mitigation, and recovery phases.

The generated ontology was validated in Protégé and deployed as the knowledge base for a DRM chatbot integrated with Large Language Models (LLMs) and Retrieval-Augmented Generation (RAG) to enhance contextual response generation and information retrieval.

Biography: Dr. Jennifer Contreras is the founder of JOCTech ICT Consultancy and Training Services and a faculty member in the Department of Software Technology at De La Salle University - Manila. With over two decades of experience in academia, she had the privilege of teaching and connecting with students from diverse backgrounds across the globe (Libya, Sudan, Malaysia, Sweden, and the Philippines). This journey has exposed her to a rich array of beliefs, cultures, and educational backgrounds among her students.
She earned her Ph.D. in Information and Communication Technology from Al Madinah International University in Malaysia with the dissertation study entitled, “Automated Question Generator for Automated Essay Scoring System using Ontology and Natural Language Processing”. Now, she is currently finishing her Ph.D. in Computer Science, a dissertation study at the Technological Institute of the Philippines-Manila entitled, “Hybrid Support Vector Machine–Transformer Deep Learning Kernel Approach for Ontology Learning in Disaster Risk Management”. Her research interests focus on Artificial Intelligence, Machine Learning, Natural Language Processing, Data Mining, and Data science. Within these domains, she has already published various papers in collaboration with different researchers from both local and international backgrounds. Her solid foundation was further enhanced through multi-disciplinary research when she took charge as Director at the Innovation and Research Development Center, where she managed various projects and initiatives to drive innovation and advance the organization's research goals. Dr. Contreras has also served as Faculty Member at Linnaeus University (Sweden) and Al-Madinah International University, and previously as Project Development Officer IV at the Department of Information and Communications Technology (DICT), where she led MOOC and LMS initiatives for national digital upskilling. Her current DOST-MIRDC funded research focuses on predicting thermal properties of 3D-printed materials through data mining to promote sustainable manufacturing. Widely published in IEEE and international journals, she advocates for the ethical and inclusive use of AI in education, research, and public service, empowering learners to thrive in a technology-driven world.