Prof. David Thomas, University of Southampton
Prof. David Thomas studied Computer Science as an undergrad at the Dept. of Computing in Imperial, then he did his PhD in digital architectures in the same department. After 5 years as
a researcher associate and then research fellow, in 2010 he moved to the Dept. of Electrical and Electronic Engineer at Imperial as a Lecturer, then Senior Lecturer. In 2021 he joined the Electronics and Computer Science Dept. as a Professor. Both his research and teaching interests are at the intersection of software and hardware, particularly in the interaction and relationshops between programming languages, algorithms, computer achitecture and digital implementation. A lot of his research involves the use of FPGAs (Field Programmable Gate Arrays), as they provide a great playground for exploring and implementing new digital architectures, such as custom CPUs, application-specific accelerators, or new programming paradigms such as event-driven computing.
Time of talk: Tuesday 13 October, 09:10-10:10
Title: Explainable AI-assisted risk modelling for national financial markets
Abstract:
Machine learning and “AI” has been very successful in providing black-box predictions and inferences in a wide variety of fields, including analysis of financial markets. If we only want to make a prediction about the future price path of an asset, or to create a buy/sell decision for a portfolio, then there are a lot of naïve and sophisticated methods that do “quite well” – or at least well enough to turn a profit in a commercial setting. But at a governmental and regulatory level this is not enough, as decision makers don’t only need predictions on what markets are likely to do, they also need to understand where the predictions came from, and what the underlying data and knowledge supporting that prediction is. Within a project called SONNETS we are trying to tackle this problem of addressing national financial risk prediction, while also providing explanations of risk assessments needed by decision makers such as the Bank of England. This talk will present an overview of the SONNETS approach, and then focus in on techniques we have found particularly useful. This includes direct use of Tsetlin Machines for prediction, but also as a supporting mechanism for explaining where predictions come from.
Dr. Mirza Khalid Baig, National Institute of Technology, Rourkela, Odisha, India
Dr. Mirza Khalid Baig received his Ph.D. degree from the Centre for Bio-Inspired Technology (CBIT), Department of Electrical Electronics Engineering, Imperial College London, London, U.K., in 2018, under the supervision of Prof. Christofer Toumazou. After his PhD, he was a Postdoctoral Research Associate with Centre for Bioinspired Technology, Imperial College London. His work at Imperial was funded through the European Research Council (ERC) and the Engineering and Physical Sciences Research Council (EPSRC) Grants. Before working in academia, he worked in the Industry as an Electronics Engineer in the Product Design Team, to implement a novel authentication technology called Laser Surface Authentication (LSA).
He is currently an Assistant Professor at the Department of Biotechnology and Medical Engineering, National Institute of Technology, Rourkela. His current research interests include developing AI-driven, wearable, biosensors and closed-loop medical devices for healthcare and precision medicine applications. His research at NIT Rourkela has been funded by Department of Science and Technology (DST), Indian Council of Medical Research (ICMR), Department of Biotechnology (DBT) and Aushandhan National Research Foundation (ANRF).
Time of talk: Tuesday 13 October, 17:00-18:00
Title: Tsetlinomics: Simplifying and interpreting multi-omics data using Tsetlin machines
Abstract:
Multiomics data are rich in terms of the information they carry and convey. However, their information content is overshadowed by the inherent data complexity. This makes it difficult to derive inferences from the omics dataset without use of complex bioinformatics tools. In this talk, we will be looking at understanding how the simple interpretable logic rule approach of Tsetlin machines can be used to handle the complexity of omics datasets. The talk will cover the information theoretic aspect of multi-omics data and its link to Tsetlin machines. The talk will demonstrate the application of Tsetlin machines in developing single disease and cross disease risk estimations, discovering hidden patterns in multi-omics datasets, discovering genetic, epigenetic and metabolomic biomarkers. Finally, we will look toward the future of decentralized medicine, showcasing how the low-power, bitwise architecture of Tsetlin machines enables Edge AI deployment directly onto point-of-care diagnostic devices.
Dr. Gouthaman K V is a Researcher at Dolby Laboratories, working on foundation AI models for multi-modal understanding across audio, speech, music, video, and text. His research spans multi-modal AI, with a strong focus on audio-centric intelligence for next-generation media technologies. At Dolby, he also works on perceptual audio quality modeling, music and speech enhancement and restoration, as well as trustworthy, explainable, and responsible AI for multimedia applications. Prior to Dolby, he was a Senior Data Scientist at Myntra (Flipkart/Walmart), where he worked on multi-modal search and language-based retrieval systems for large-scale e-commerce applications. He received his PhD in multi-modal AI from IIT Madras, where his research focused on vision-language models, including visual question answering, image and video captioning, and related problems in multi-modal learning. Earlier, he worked on classical computer vision problems such as object tracking and facial expression recognition. He also actively contributes to the research community and serves as a reviewer for leading venues including CVPR, ICCV, ECCV, NeurIPS, ICASSP, and TPAMI etc.
Dr. Aveen Dayal is a Senior Multimodal AI Researcher at Dolby Advanced Technology Group India. He received his Ph.D. in Artificial Intelligence from the Indian Institute of Technology Hyderabad, where he was a Prime Minister’s Research Fellow and received the IIT Hyderabad Excellence in Research Award 2024. His research spans multimodal and generative AI, out-of-distribution robustness, efficient inference, content protection and attribution, and model evaluation. His work has been published at leading venues, including CVPR, WACV, ECCV, NeurIPS, and IEEE Transactions on Image Processing. Aveen has previously worked with Adobe Research and Microsoft Research India on generative AI, multimodal models, and efficient large-language-model inference. He has also been a visiting researcher at the University of Agder, Norway, where he contributed to interdisciplinary projects involving autonomous systems, biomedical imaging, sensor-based perception, and edge AI. He regularly contributes to the research community as a reviewer for major AI conferences and journals.
Dr. Gouthaman K V and Dr. Aveen Dayal will do a combined talk, and abstract and title is presented below.
Time of talk: Wednesday 14 October, 09:10-10:10
Title: Towards Trustworthy AI: From Internal Representations to Reliable Behaviour
Abstract:
Recent advances in foundation models have transformed artificial intelligence across language, vision, audio, and multimodal applications. As these systems become increasingly capable and widely deployed, understanding how they process information, make decisions, and produce reliable outputs has become a central challenge for trustworthy AI. This talk will provide a broad overview of emerging approaches for understanding and evaluating modern AI systems from complementary perspectives. We will discuss methods for analysing the internal representations learned by foundation models, approaches for relating these representations to model behaviour, and recent developments in understanding the reliability of model predictions. The talk will also explore key challenges such as hallucination, uncertainty, refusal behaviour, robustness, and safety in both language and multimodal AI systems, highlighting how these phenomena influence the trustworthiness of AI in real-world applications. By bringing together these perspectives, the talk will highlight current challenges and emerging directions toward building more interpretable, reliable, and trustworthy AI systems.
Tutorials
Time of talk: Monday 12 October, 14:00-15:00
Tutorial Part 1: Introduction to Tsetlin Machines: Consepts and Foundations
Tutorial abstract:
This tutorial provides an introduction to Tsetlin Machines, a novel machine learning approach based on propositional logic and learning automata. Participants will learn the fundamental concepts behind Tsetlin Machines, including their architecture, learning mechanism, and interpretability features. The tutorial will cover key variants of the Tsetlin Machine and discuss their applications across a range of domains.
The tutorial is divided into two parts. The first part introduces the theoretical foundations and core principles of Tsetlin Machines. The second part offers a hands-on session where participants will explore practical implementations, learn how to train and evaluate Tsetlin Machine models, and gain experience using available software tools and frameworks.
No prior knowledge of Tsetlin Machines is required, making the tutorial suitable for students, researchers, and practitioners interested in interpretable and efficient machine learning.
Tutorial Presenter: Prof. Ole-Christoffer Granmo
Bio Prof. Ole-Christoffer Granmo:
Prof. Ole-Christoffer Granmo is the Founding Director of the Centre for Artificial Intelligence Research (CAIR) at the University of Agder, Norway. He obtained his master’s degree in 1999 and his PhD degree in 2004, both from the University of Oslo, Norway. In 2018, he created the Tsetlin machine, for which he was awarded the AI research paper of the decade by the Norwegian Artificial Intelligence Consortium (NORA) in 2022. Dr. Granmo has authored and co-authored 195+ refereed papers with nine paper awards in machine learning, encompassing learning automata, bandit algorithms, Tsetlin machines, Bayesian reasoning, reinforcement learning, and computational linguistics. He has further coordinated 7+ research projects and graduated 55+ master- and ten PhD students. Dr. Granmo is also a Guest Professor at the Microsystems AI (MAI) Lab of Newcastle University, UK, and a co-founder of NORA. Apart from his academic endeavors, he co-founded Anzyz Technologies AS and is the Chair of the Technical Steering Committee at Literal Labs.
Time of talk: Monday 12 October, 15:30-16:30
Tutorial Part 2: Introduction to Tsetlin Machines: Practical Exercises and Coding
Tutorial abstract:
This hands-on session provides participants with practical experience implementing and training Tsetlin Machines for image classification using the MNIST handwritten digit dataset. Building on the theoretical foundations covered in the first part of the tutorial, participants will work through guided exercises using publicly available software frameworks to implement three variants of the Tsetlin Machine: the standard Tsetlin Machine (TM), the Convolutional TM (CTM), and the Coalesced TM (CoTM).
Participants will learn how to train each model variant on MNIST, and evaluate classification performance in terms of accuracy, training time, and interpretability. The session will highlight practical differences between the variants, including how the CTM captures spatial patterns through patch-based clause evaluation, and how the CoTM improves parameter efficiency by sharing clauses across classes. Participants will also examine the learned clauses to understand how TM arrive at interpretable, human-readable classification decisions.
By the end of the session, participants will have hands-on experience setting up, training, and evaluating TM models, along with a working notebook they can adapt for their own classification tasks. Basic familiarity with Python is recommended; no prior experience with TM is required.
Tutorial Presenter: Prof. Linga Reddy Cenkeramaddi, Prof. Srinivas Bobbu, Bethi Pardhasaradhi, Thrishank Hesaraghatta Shivakumar and K Koushik
Bio Prof. Linga Reddy Cenkeramaddi:
Prof. Linga Reddy Cenkeramaddi (Senior Member, IEEE) received the master’s degree in electrical engineering from Indian Institute of Technology Delhi (IIT Delhi), New Delhi, India, in 2004, and the Ph.D. degree in electrical engineering from the Norwegian University of Science and Technology (NTNU), Trondheim, Norway, in 2011. He was with Texas Instruments on mixed-signal circuit design before joining the Ph.D. Program with NTNU. After finishing the Ph.D. degree, he worked on radiation imaging for an atmosphere–space interaction monitor (ASIM mission to the International Space Station) at the University of Bergen, Bergen, Norway, from 2010 to 2012. He is currently the Leader of the Autonomous and Cyber-Physical Systems (ACPS) Research Group and a professor with the University of Agder, Grimstad, Norway. He has co-authored over 300 research publications that have been published in prestigious international journals and standard conferences in the research areas of the Internet of Things (IoT), cyber-physical systems, autonomous systems, robotics and automation involving advanced sensor systems, computer vision, thermal imaging, LiDAR imaging, radar imaging, wireless sensor networks, smart electronic systems, advanced machine learning techniques, and connected autonomous systems, including drones/unmanned aerial vehicles (UAVs), unmanned ground vehicles (UGVs), unmanned underwater systems (UUSs), 5G- (and beyond) enabled autonomous vehicles, and socio-technical systems, like urban transportation systems, smart agriculture, and smart cities. Dr. Cenkeramaddi is a member of ACM and the editorial boards of various international journals and the technical program committees of several IEEE conferences. He is the principal investigator and a co-principal investigator of many research grants from the Norwegian Research Council.
Bio Prof. Srinivas Boppu is currently working as an Associate Professor in the School of Electrical and Computer Sciences, IIT Bhubaneswar. He holds an international degree in M.Sc. (IC Design) jointly offered by Nanyang Technological University, Singapore, and the Technical University of Munich, Germany. He received his Ph.D. degree from the chair for hardware/software co-design at the Department of Computer Science, University of Erlangen-Nuremberg, Germany, in 2015. Before moving to India, he worked as a senior consultant at Infineon Technologies in Munich, Germany. He also worked with Freescale Semiconductors India and ST Microelectronics as a physical design engineer before pursuing his Ph.D. His research interests include high-level synthesis, programmable hardware accelerators, compilers, scheduling and mapping approaches, low-power VLSI design, SoC design, and design automation of integrated circuits. He has 20+ years of experience in both academia and industry in the field of the VLSI domain. He was awarded a full scholarship by Infineon Technologies Asia Pacific Pte. Ltd. for M.Sc. (IC Design) jointly offered by NTU, Singapore, and TUM, Germany.
Bio Dr. Bethi Pardhasaradhi
Dr. Bethi Pardhasaradhi received his B.Tech degree in Electronics and Communication Engineering from Jawaharlal Nehru Technological University Kakinada (JNTU-K), Andhra Pradesh, India, in 2014. Completed M.Tech degree in VLSI design from ABV-Indian Institute of Information Technology and Management (ABV-IIITM), Madhya Pradesh, India, in 2016. He received Ph.D. degree in Electronics and Communication Engineering from the National Institute of Technology Karnataka (NIT-K), India, 2021. He was a Visiting Ph.D. Scholar for 16 months at Estimation Tracking and Fusion (ETF) Laboratory, McMaster University, Canada, under the supervision of Prof. T. Kirubarajan, during 2018-2019. In addition, he was also a Visiting Researcher as part of Indo-Norwegian Collaboration to Autonomous and cyber-physical systems (ACPS) research group, department of Information and Communication Technology (ICT), University of Agder, Grimstad, Norway, under the supervision of Prof. Linga Reddy Cenkeramaddi, during 2021-2022. After his Ph.D, he worked for two years as a technical Specialist in the ADAS Department in Continental Autonomous Mobility India Pvt Ltd, India. Currently working as a postdoc fellow in ICT dept from 2022 in university of Agder (UiA), Norway.
He was a recipient of Sir C. V. Raman Award from the Institution of Engineering and Technology (IET) for Outstanding Academics and Research. Received Best Ph.D Thesis 2022 in Graduate Thesis Evaluation in 7- minutes (GraTE-7) from IEEE ComSoc Graduate Congress. In addition, he received IEEE ITS research excellence award in 2023 from IEEE information theory society (ITS) Bangalore chapter. Moreover, receipt of Protsahan award from IEEE ComSoc and best papers recognition from IEEE conferences. His research interests include intentional interference to autonomous sensors, target tracking, information fusion, and explainable AI.
Bio Thrishank Hesaraghatta Shivakumar
Thrishank Hesaraghatta Shivakumar is currently pursuing his PhD, jointly offered by the Indian Institute of Technology Bhubaneswar, India, and the University of Agder, Norway. His doctoral research, “Energy Efficient Hardware and Software Design for Explainable Edge AI Applications,” focuses on hardware-accelerated Tsetlin Machines for autonomous multi-sensor data processing, spanning real-time low-energy classification, FPGA-based hardware acceleration, and real-time prototyping for edge AI systems. He holds a Master’s degree from Nitte Meenakshi Institute of Technology, Bengaluru, India, during which he served as a Project Trainee at the Laboratory of Electro-Optics Systems, Indian Space Research Organisation (LEOS-ISRO), developing complete prototypes of Alpha- and Gamma-Dosimeters for his master’s thesis. His research interests include Tsetlin Machines, hardware architectures, and energy-efficient accelerators for artificial intelligence. He is presently working on multi-sensor data and multi-sensor fusion using Tsetlin Machine composites.
Bio K Koushik:
K Koushik is currently pursuing an integrated Master’s and Ph.D. jointly at the Indian Institute of Technology (IIT) Bhubaneswar, India, and the University of Agder (UiA), Norway. He defended his Master’s thesis, “Complementary Sensing-Based Adaptive Radar Target Tracking,” at UiA in December 2025. He was also awarded the Best Student Runner-Up Award for the paper “PruneTM: Clause-Driven Feature Ranking and Pruning in Tsetlin Machine” at the International Symposium on Tsetlin Machines, Rome, Italy, in 2025. His current research focuses on designing interpretable, energy-efficient, and low-latency algorithms for multimodal data processing on edge devices.