Operationalizing AI : Safe, Efficient, and Privacy-preserving AI Systems at Scale

This session focuses on the intrinsic safety of artificial intelligence systems and their ability to behave in a predictable and controllable
manner, in line with human intentions, even in critical or high-risks situations. Topics will include goal alignment, human oversight, evaluation and validation methodologies, aimed at preventing harmful behaviors or unintended optimization effects.

This session addresses the challenge of developing and deploying AI Systems while ensuring the protection of personal data and fundamental rights, in a context of large-scale data use and increasing regulatory requirements. It will cover privacy-preserving machine learning techniques, such as federated learning, differential privacy, secure computation, confidential computing and data minimization strategies. The session will explore the trade-offs between privacy, performance, and scalability as well the integration of privacy by design principles into real world systems.

This session focuses on the challenge of deploying AI systems
at scale while maintaining high levels of efficiency, and accuracy. As AI models grow in size and complexity and are deployed in the real world, achieving strong performance under real-world constraints, such as latency, throughput, cost and infrastructure limitations, has become a central technical and operational issue. Topics will include algorithmic and model efficiency techniques for improving
accuracy per unit of computation, model compression and optimization, efficient training and inference pipelines, and hardware-software codesign. By emphasizing efficiency and accuracy rather than sheer model size, this session highlights how scalable AI systems can deliver precise results in industrial, scientific and public sector applications, enabling broader adoption without
compromising performance or operational feasibility.

Senior Researcher at the CNRS-AIST Joint Robotics Labotatory
Website / LinkedIn

Mehdi Benallegue received the ingénieur degree from the Institut National d’Informatique (INI), Algeria, in 2007, the M.Sc. degree from the University of Paris 7, Paris, France, in 2008 and the Ph.D. degree from Université de Montpellier 2, France, in 2011. He has been a postdoctoral researcher in a neurophysiology laboratory in Collège de France and in LAAS CNRS. He is currently a Research Associate with the CNRS-AIST Joint Robotics Laboratory in National Institute of Advanced Industrial Science and Technology (AIST), Japan. His research interests include estimation and control of robots, biomechanics, neuroscience and computational geometry.

Director of the Department of Computer Science and Information Technology, National Institute of Applied Sciences of Lyon, France
Website / LinkedIn

Lionel Brunie is full professor at the National Institute of Applied Sciences (INSA) of Lyon, France.

Lionel Brunie is director of INSA department of computer science and information technology (~500 students, 45 faculties and staff), a European leading academic institution training high level CS&IT engineers.

Along with Pr. Harald Kosch (University of Passau, Germany) and Pr. Ernesto Damiani (University of Milan, Italy), Lionel Brunie has headed the IRIXYS Center (International Research and Innovation Centre in Intelligent Digital Systems), since its creation in 2016. IRIXYS provides a wide range of services in research, innovation, and education.

He founded and he has co-chaired with Pr. H. Kosch an international Franco-German Chair of Excellence “AI and Cybersecurity” since 2026.

Sorbonne University / CNRS / University of Tokyo
Website

Philippe Codognet is the director of the Japanese-French Laboratory for Informatics (JFLI), a joint laboratory between CNRS (French National Center for Scientific Research), Sorbonne University (Paris), University of Tokyo, National Institute of Informatics and Keio University.
He received a Ph.D. in Computer Science from University of Bordeaux-I (France) in 1989, and was researcher at INRIA, with a sabbatical at Sony Computer Science Laboratory (Paris), before becoming professor at Sorbonne University in 1998. He also worked as attaché for science and technology at the French Embassy in Japan (Tokyo) and at the French Embassy in Singapore.

Professor at RPTU, Executive Director at DFKI
Website / LinkedIn

Andreas Dengel is Professor for Artificial Intelligence at RPTU Kaiserslautern-Landau since 1993 and Executive Director of the German Research Center for Artificial Intelligence (DFKI) in Kaiserslautern. Since 2009, he has also been a Professor (kyakuin) at the Department of Computer Science and Intelligent Systems at Osaka Metropolitan University (OMU). Andreas published more than 800 peer-reviewed research papers, founded a dozen of start-ups and received many awards, e.g., he is a recipient of the most prestigious Order of Merit of the State of Rhineland-Palatinate and of the oldest Japanese order, the “Order of the Rising Sun, Gold Rays with Neck Ribbon”.

AI Scientist, Code Generation Team, Mistral AI
Website / LinkedIn

Engineering diploma from Ecole Polytechnique, France. Started working at mistral two years ago in code and reinforcement learning, with contributions in both science and infra. Participated in the release of several flagship models of Mistral (Magistral, MS4, MM3.5).

Senior Fellow of CEA (France)
Website / LinkedIn

Dr. Marc Duranton is Senior Fellow of CEA and member of the Digital Systems and Integrated Circuits Division of CEA, where he is involved in realisations (hardware accelerators and software tools) for Artificial Intelligence and for distributed systems from IoT to HPC.
He previously worked in Philips and NXP where he led the development of the family of L-Neuro chips, digital accelerators for artificial neural networks and on several video coprocessors for the VLIW processor TriMedia.
He is in charge of the roadmap activities of the HiPEAC community.

Head of Institute for Artificial Intelligence in Cardiovascular Medicine, Heidelberg University Hospital
Website / LinkedIn

Prof. Dr. Sandy Engelhardt is the Head of the Institute for Artificial Intelligence in Cardiovascular Medicine at Heidelberg University Hospital. Her research focuses on medical image processing, computer-assisted surgery, and the development of artificial intelligence methods for cardiovascular applications. She leads an interdisciplinary team dedicated to bridging the gap between computer science and clinical practice.

Minister Counsellor for Digital Economy Policy, Delegation of the EU to Japan
Website / LinkedIn

Peter FATELNIG is Minister Counsellor for Digital Economy at the Delegation of the European Union to Japan. Digital Economy is a top-priority for the European Union and he leads a team implementing the EU-Japan Digital Partnership. Peter is committed to a positive European vision for an economy and society taking full advantage of digital technologies in a fair and secure way.

Prior to his current role, from 2018 to 2023, Peter managed the digital economy portfolio at the Delegation of the European Union to the United States. Before joining the European External Action Service, Peter was a senior manager at the European Commission where he started in 1998. Peter began his international career by working on assignments for the advisory firm American Management Systems, and for the European Space Agency, in the Netherlands. Peter holds a Master degree in Communication Engineering from the University of Technology in Graz, Austria.

Senior Researcher, German Research Center for Artificial Intelligence (DFKI)
Website / LinkedIn

He is a researcher at the German Research Center for Artificial Intelligence with over six years of experience in foundational research on generative AI. His work focuses on multimodal generative models, controllable image synthesis, efficient vision and generation algorithms, super-resolution, dataset distillation, and continual learning. He received his Ph.D. in Computer Science from RPTU Kaiserslautern-Landau in 2025, advised by Prof. Andreas Dengel, with a thesis on Controllable Deep Image Synthesis. He has authored over 35 publications with more than 800 citations and previously interned at Adobe Research and Meta AI.

Professor for Media and Culture Theory / Vice-President International University of Europe for Applied Sciences
Website / LinkedIn

Professor of Media and Cultural Theory and Vice President for International Affairs and Institutional Development. Her research investigates how artificial intelligence, robots and digital infrastructures are culturally imagined, aesthetically staged and epistemologically contested across gender, race and geopolitical difference. Drawing on posthumanist media theory, Science and Technology Studies and Indigenous epistemologies, she develops cross-cultural and critical frameworks for thinking AI, technology ethics and future-making. She has initiated and led international collaborations on desirable AI and critical media culture.

President of the Franco-German University
Website / LinkedIn

Prof. Dr. Philippe GRECIANO is Full professor, Jean Monnet Chair of Excellence, and President of the Franco-German University. He has initiated several tri-national partnerships with Japan.

Managing Director, Enobyte GmbH
Website / LinkedIn

Dr. Hermann Gumpp studied physics and computer science at LMU Munich and earned a PhD in biophysics. In Japan, he worked as a consultant for a German-Japanese corporate group and at the National Institute for Informatics (NII) in Tokyo. He serves on the board of the German-Japanese Society (DJG) in Bavaria and leads the IT working group of the German-Japanese Business Association (DJW). As Managing Director of Enobyte GmbH in Munich, he and his team of data protection officers and IT experts support German and Japanese companies in implementing GDPR compliance and cybersecurity. He is co-author of the Japanese Amazon bestseller “GDPR Guidebook.” Since 2021, he has been a volunteer Special Advisor for Cybersecurity & Data Protection at DJW.

Team lead for data and service ecosystems at German Aerospace Center (DLR)
Website / LinkedIn

Svenja Hainz is leading the team on Data and Service Ecosystems at DLR’s institute for AI Safety and Security. In this capacity she is coordinating and supporting the researchers who are working on creating fair and secure data spaces for different sectors.

In 2017, she joined DLR as a scientific researcher at the Institute of Transportation Systems with her main focus on the integrated impact assessment of railways systems. Afterwards she was DLR’s representative at European level in the field of transportation research.

She holds degrees in transportation planning and operation and in Industrial Engineering and Management.

Principal Researcher, National Institute of Advanced Industrial Science and Technology (AIST)
Website

Goichiro Hanaoka received his Ph.D. from the University of Tokyo in 2002 and joined AIST in 2005. He is currently Principal Researcher at the Cyber Physical Security Research Institute, AIST. His research interests include cryptography, privacy-enhancing technologies, and trustworthy data utilization for data-driven society. He has been promoting the practical deployment of advanced security technologies and studying how privacy and usability can be balanced in real-world systems. He received the Commendation for Science and Technology from the Minister of Education, Culture, Sports, Science and Technology in 2018.

Secretary General, Tokyo Centre of the GPAI Expert Community
Website

Dr. Yuko Harayama is Secretary General of the Tokyo Centre of the Global Partnership on AI (GPAI) Expert Community, Trustee of Yamaguchi University, and Board Member of Toray Industries. Her career spans government, international organizations, and academia. She has served as RIKEN’s Executive Director for International Affairs, an Executive Member of Japan’s Council for Science, Technology and Innovation in the Cabinet Office, and Deputy Director of the OECD’s Directorate for Science, Technology and Industry. Earlier, she was a Professor at Tohoku University’s Graduate School of Engineering for a decade.

Research group lead Intelligent Software Systems – Institute of Software Technology – German Aerospace Center (DLR)
Website / LinkedIn

Tobias Hecking received his Ph.D. degrees in computer science from the University of Duisburg-Essen, in 2016. Since 2020, he has been the leading the Intelligent Software Systems group of the Institute of Software Technology at German Aerospace Center (DLR). In his research he combines methods of complex network analysis, natural language processing, and machine learning for aerospace applications and beyond. He is active in several national and international research projects in the field. Furthermore, Tobias Hecking is leading the DLR’s research initiative to advance generative AI in various aerospace applications and coordinates the development of the corresponding technology stack.

Researcher, NEC Corporation / Intent Exchange, Inc.
LinkedIn

Florence Ho is a researcher at NEC Corporation, Tokyo, where she is currently a member of the Knowledge Science Research Lab. She is also a researcher at Intent Exchange, Inc.
Previously, she worked as a researcher at the National Institute of Advanced Industrial Science and Technology (AIST), Japan.
She received her PhD in Informatics from the Graduate University for Advanced Studies, SOKENDAI, Japan, in 2020. She also holds an MSc in Applied Mathematics from Pantheon-Sorbonne University, Paris, and an engineering degree from INP-ENSEEIHT, Toulouse.
Her research interests include optimization, multi-agent systems, and intelligent transportation systems.

Professor, KAIST
Website / LinkedIn

Sung Ju Hwang is an Endowed Chair Professor in the Kim Jaechul Graduate School of AI at KAIST, Founder and CEO of DeepAuto.ai, an enterprise agentic AI company, and an Advisory Professor on the Presidential Committee on the AI Economy. He received his Ph.D. from the University of Texas at Austin, and previously held positions at Disney Research and UNIST. He has published over 180 papers at AI venues such as NeurIPS, ICML, ICLR and CVPR, and serves as an Associate Editor of IEEE TPAMI. His research spans self-evolving agentic AI, efficient AI, and AI for scientific discovery.

Member of CNRS-AIST Joint Research Lab, Visiting Professor at Institute of Tokyo Science
Website / LinkedIn

Kristiina Jokinen is a Member of CNRS-AIST Joint Robotics Lab and Visiting Professor at Institute of Science Tokyo. Originally from Finland, she received her PhD from Manchester, and was JSPS Fellow at NAIST. She was Senior Researcher at AIST AI Research Center and led dialogue research in the large EU-Japan collaboration project Evita. Her research concerns human-robot interaction, generative dialogue models, knowledge-graphs, and multimodal communication, and she has published widely on these topics. She co-developed WikiTalk, a Wikipedia-based robot dialogue system which won Best Robot Design award (Software Category) at ICSR in 2017.

Professor, CDS, Tohoku University
Website / LinkedIn

Minoru Kuribayashi is a Professor in Center for Data-driven Science and Artificial Intelligence (CDS) at Tohoku University. His research interests include multimedia security, digital watermarking, cryptography, and coding theory. He received the Young Professionals Award from IEEE Kansai Section in 2014. He is appointed as a distinguished lecturer of APSIPA (Asia-Pacific Signal and Information Processing Association) in 2025-2026.

Privacy by Design Lab
Website / LinkedIn

Kohei is the Co-Founder of Privacy by Design Lab, a non-profit dedicated to driving data privacy culture and policymaking. He collaborates globally with governments, businesses, and international watchdogs to advance fundamental privacy rights. A recognized expert in data privacy and blockchain, he contributes to open-source projects and has spoken at major forums like UNESCO. Additionally, he leverages his extensive non-profit and education experience to help politicians worldwide develop impactful public policy.

Director, Korea AISI (AI Safety Institute)
Website / LinkedIn

Dr. Myuhng-Joo Kim is the founding Director of the Korea AI Safety Institute (AISI), established in November 2024. He earned his BS, MS, and PhD in Computer Science from Seoul National University and served as a professor at Seoul Women’s University for three decades. Trained in computer engineering, he has long been engaged in information security and digital ethics. He developed Korea’s first AI ethics framework, Seoul PACT, earning the Geunjeong Order of Service Merit, and authored AI Has No Conscience (2022). He also serves on the Supreme Court AI Committee and as an expert with the OECD GPAI (Global Partnership on AI).

University of Duisburg-Essen, Research Center Trustworthy Data Science and Secuity
Website

Dr. Nicole Krämer is Full Professor of Social Psychology, Media and Communication at the University of Duisburg-Essen, Germany, and Director of the Research Center “Trustworthy Data Science and Security”. She completed her PhD in Psychology at the University of Cologne in 2001. Dr. Krämer has been pioneering interdisciplinary research on human-technology-interaction and opinion building in social media for 20 years. Empirical research on human-AI-interaction as well as disinformation is an important focus of her work. She served as Editor-in-Chief of the Journal of Media Psychology and currently is Associate Editor of the Journal of Computer Mediated Communication.

Tenured Researcher (ISFP) at Inria
Website / LinkedIn

Mohamed Maouche is a researcher at Inria Lyon (ISFP) and member of the PRIVATICS team since 2022. His interest is to build machine learning systems that manage a good trade-off between privacy and utility. This includes working on data anonymization and privacy preserving machine learning. Previously, he was a one-year post-doc whithin the Chaire DSVD supported by Renault and Labex IMU working on privacy issues in Federated Learning. He also spent two years as a post-doc in MAGNET Team (Inria Lille) to work on speech anonymization. He defended his PhD from Insa Lyon in 2019 on the subject of Location privacy.

Advisor the Chairman, Sakana AI
Website / LinkedIn

Matei Mihalca serves as Advisor to the Chairman of Sakana AI. He is the co-founder of Forward Partners, a spin-off from the Center of Research Enrico Fermi (CREF) in Rome. Matei began his career as the Taiwan semiconductors analyst at Merrill Lynch and subsequently served in a variety of roles at Goldman Sachs, AQR Capital, Macquarie Group, and CITIC Capital. Matei was a PhD and MA student at Harvard University, and an exchange student in Beijing, China.

Senior Researcher, National Institute of Information and Communications Technology
Website

Her research is on the philosophical and ethical dimensions of the new circumstances that such technologies will bring about (Ethical, Legal, and Social Issues: ELSI). More broadly, her area of specialty is philosophy of mind. In particular, she works on issues such as self-control, rationality, and moral responsibility, focusing on the role of emotions, using methods based in analytic philosophy. Her approach incorporates the results of cognitive sciences such as psychology and neuroscience in addition to traditional philosophical perspectives when examining various philosophical topics.

RIKEN Center for Advanced Intelligence Project
Website / LinkedIn

Takayuki Osa is Team Director of the Robot Learning Team at the RIKEN Center for Advanced Intelligence Project (AIP), Japan. His research focuses on efficient and reliable robot learning for scalable physical AI. Prior to joining RIKEN, he served as Associate Professor at the University of Tokyo and Kyushu Institute of Technology. He received his Ph.D. in Engineering from the University of Tokyo in 2015 and was a postdoctoral researcher at TU Darmstadt, Germany from 2015 to 2017.

Associate Professor, University of Tokyo
Website / LinkedIn

Grant Jun Otsuki is a cultural anthropologist who specializes in studies of technology in Japan. He holds a Ph.D. in anthropology (U. Toronto) and a M.S. in Science and Technology Studies (RPI, USA). Grant conducts ethnographic and historical research on Japanese scientists and engineers involved in computing, human-machine interface research, and artificial intelligence to understand the cultural dimensions of digital technologies. He has previously held positions at the University of Tsukuba and Victoria University of Wellington (New Zealand.)

Deputy Head – DFKI-MAP German Research Center of Artificial Intelligence DFKI
Website 1 / Website 2 / LinkedIn

Dr.-Ing. Alexandra Pehlken has been Deputy Head of the Marine Perception (MAP) research division at DFKI Oldenburg since October 2025. From 2018 to 2025, she worked at OFFIS – Institute for Computer Science as Group Leader for Sustainable Manufacturing. As a renowned expert in raw materials and sustainability, she develops digital tools (with and without AI) to conserve resources. Alexandra Pehlken places particular emphasis on the issue of critical raw materials, with the aim of promoting sustainable practices. She completed her PhD at RWTH Aachen University and spent 23 months as a postdoctoral researcher at Natural Resources Canada (NRCan) in Ottawa.

Research Scientist, Inria and Senior Research Fellow, University of Oxford
Website / LinkedIn

Carina Prunkl is a Research Scientist at Inria’s REGALIA project and a Senior Research Fellow at the University of Oxford’s Institute for Ethics in AI. She is also Lead Writer of the 2026 International AI Safety Report and Co-Investigator of the UKRI-funded project CHAILD. Her research focuses on agency, autonomy, and human oversight, as well as evaluation standards and the governance of AI more broadly. Carina holds a BSc and MSc in Physics from Freie Universität Berlin, an MSt in Philosophy of Physics, and a DPhil in Philosophy from the University of Oxford. She was previously a postdoctoral researcher at the Future of Humanity Institute and the Institute for Ethics in AI at Oxford, and an Assistant Professor at Utrecht University.

Keio University School of Medicine
Website

In his quest to understand life, consciousness, evolution, disease and ageing, Professor Kazuhiro Sakurada has conducted research in industry and academia. In 2008, he shifted his focus from regenerative medicine and drug discovery to AI medical data science. There, he developed key technology for disease foundation models and organism mechanical theories. His book The Origin of Subspecies, published in September 2020, was highly acclaimed for offering a novel perspective on life.

Project Professor, National Institute of Informatics / Deputy Director, AI Safety Institute
Website

Satoshi Sekine is a Project Professor, National Institute of Informatics and Deputy Director, AI Safety Institute. He received PhD at New York University in 1998. He has been working on Natural Language Processing, in particular Information Extraction, Named Entity, Question Answering and other related topics. Since 2022, he has been working on building datasets for training and evaluating Japanese LLM datasets, i.e. general instruction data and safety data. Currently, he is leading JAI-Trust benchmark data for measuring AI safety as a community efforts with 100 researchers.

University of Lille & Inria (French Institute of Computer Science)
Website / LinkedIn

Juliette Sénéchal is Professor of private law, specialised in Digital Law, at the University of Lille, and Deputy Scientific Director of Inria (French Institute of Computer Science) in charge of Humanities and Social Sciences. She is also member of the european Alliance Neurotech.EU. Her current topics of research focus on legal aspects of individual and collective impacts of artificial intelligence, digital services and neurotechnologies.

Director, RIKEN Center for Advanced Intelligence Project / Professor, The University of Tokyo
Website

Masashi Sugiyama received his Ph.D. in Computer Science from Tokyo Institute of Technology, Japan, in 2001. After serving as an assistant and associate professor at the same institute, he became a professor at the University of Tokyo in 2014. Since 2016, he has also served as the director of the RIKEN Center for Advanced Intelligence Project. His research interests include theories and algorithms of machine learning. He was awarded the Japan Academy Medal in 2017 and the Commendation for Science and Technology by the Minister of Education, Culture, Sports, Science and Technology of Japan in 2022.

Professor, Tohoku University
Website

Jun Suzuki is a Professor at the Graduate School of Information Sciences and Director of the Center for Language AI Research at Tohoku University. Before joining Tohoku University, he was a Researcher and later a Distinguished Researcher at NTT Communication Science Laboratories. He received his Ph.D. in Engineering from the Nara Institute of Science and Technology in 2005. He was previously a Visiting Researcher at MIT CSAIL and Google, and currently serves as a Visiting Researcher at RIKEN AIP and a Visiting Professor at the National Institute of Informatics. His research focuses on natural language processing, machine learning, and AI.

CEO, AiCAN Inc.
Website / LinkedIn

Dr. Kota Takaoka is Founder & CEO of AiCAN Inc., a GovTech startup spun out of Japan’s National Institute of Advanced Industrial Science and Technology (AIST), building AI-powered SaaS platforms for child welfare institutions across around 30 municipalities in Japan. He holds a doctorate in education and is a licensed clinical psychologist and certified forensic interviewer, with direct casework experience in child abuse response divisions at Tokyo’s child welfare agencies. Dr. Takaoka is committed to AiCAN’s vision: a safe world for every child.

Professor of Nature-Inspired AI for Science at Leibniz Universität Hannover and TIB – Leibniz Information Centre for Science and Technology
Website / LinkedIn

Sahar Vahdati is Professor of Nature-Inspired AI for Science at Leibniz Universität Hannover and TIB – Leibniz Information Centre for Science and Technology. Her research explores nature-inspired intelligence for science, focusing on adaptive reasoning, abstraction, human cognition, and knowledge infrastructures. She develops AI systems that learn efficiently, discover rules, reason over scientific knowledge, and collaborate with humans in scientific workflows. Her work connects ARC-AGI, knowledge graphs, LLM reasoning, and science integrity to build trustworthy AI for discovery and scholarly communication.

Research Director at CNRS-Inria
Website / LinkedIn

Serena Villata is a research director at the CNRS within the I3S Laboratory. She has held a chair at the Institut 3IA Côte d’Azur since 2019, has served as the institute’s deputy scientific director since 2021, and has been the scientific director of the IA Cluster since November 2025. She leads the MARIANNE research team at Inria. In November 2021, she received the “Young Researchers” Award from Inria and the French Academy of Sciences. In 2010, she earned her Ph.D. from the University of Turin (Italy) for her work on computational reasoning in Artificial Intelligence. Her research field is artificial intelligence (AI), and her current work focuses on computational argumentation, with a particular emphasis on legal and medical texts, political debates, and harmful content on social media (abusive language, misinformation). Her work combines argumentation-based reasoning frameworks with natural language arguments extracted from texts. She is the Principal Investigator (PI) of the ERC Consolidator Project PANDORA, which begins in June 2026.

Dr. André Greiner-Petter
Postdoc, Institute of Computer Science, University of Goettingen, Goettingen, Germany

Poster Title:
AI for Science: Do You Prove What You Claim?

Poster Abstract:
The “AI for Science” paradigm is accelerating global research, yet its success relies entirely on the integrity of the underlying literature. As AI rapidly synthesizes discoveries, the risk of propagating unverified, hallucinated, or even plagiarized claims grows exponentially. Addressing this critical bottleneck, our collaborative French-German-Japanese research consortium introduced SciClaimEval, a benchmark for cross-modal, multi-domain scientific claim verification.

Building on the global engagement with this task, our trilateral initiative is now developing SciClaimEval 2.0. This next-generation framework will operationalize trustworthy AI by, among other things, introducing cross-document evidence, mathematical claims, and broader domain coverage. Constructing this massive, future-proof evaluation infrastructure requires extensive cross-cultural expert annotation and collaboration. Ultimately, securing funding for this expansion will establish the global standard needed to ensure tomorrow’s AI-driven breakthroughs remain verifiable, safe, and strictly grounded in evidence.

Biography:
Dr. Andre Greiner-Petter is a Postdoctoral Researcher at the University of Goettingen, based at the National Institute of Informatics (NII) in Tokyo. He holds a PhD in Engineering Sciences (Summa Cum Laude) from the University of Wuppertal and Mathematics degrees from TU Berlin.

His research in Semantic Document Analysis and Information Retrieval explores how Large Language Models comprehend complex mathematical and scientific literature. Currently, he develops robust infrastructure for “AI for Science,” focusing on automated claim verification, academic recommender systems, and cross-modal plagiarism detection. Previously, he co-founded the fintech startup Sinpex and researched at NIST.

Prof. Dr. Yosuke Senju
Associate Professor, Research Institute for Interdisciplinary Science (RIIS), Okayama University

Poster Title:
Analysis of lipid-protein interactions using AI

Poster Abstract:
Approximately 30% of the proteins encoded by the human genome interact with cell membranes. Proteins that regulate intracellular membrane systems, such as membrane proteins, are involved in various intracellular functions, including the transport of substances via the endoplasmic reticulum. Furthermore, they are involved in the pathogenesis of many serious diseases, including cancer and Alzheimer’s disease, and have become important targets in drug discovery. However, even when candidate lipid-protein interactions are experimentally identified by proteomics, the possibility of false positives due to non-specific binding cannot be ruled out, leading to vast amounts of data simply accumulating in databases. The aim of this study is to quantitatively identify reliable interaction molecules for membrane proteins by establishing a feedback loop between AI-based in silico analysis and experimental validation. This will help address various challenges in the medical field.

Biography:
Dr. Yosuke Senju received his BS in Physics from Tohoku University. He obtained his PhD in Physics from the Department of Physics, Graduate School of Science, Tohoku University, where he studied actin cytoskeleton and myosin motor protein in membrane dynamics. He studied lipid–protein interactions as a postdoc and as an assistant professor at the University of Tokyo, and as a postdoc at University of Helsinki. He studied membrane biophysics to understand lipid–protein interactions. He is currently an associate professor at the Research Institute for Interdisciplinary Science at Okayama University, studying the membrane–cytoskeleton interface using biophysics, synthetic biology, and structural biology.

Prof. Dr. Susanne Brucksch
Associate Professor, Global Japanese Studies, Teikyo University

Poster Title:
Critical AI Literacy: Paving a path to safe and competent handling of genAI tools for Global Japanese Studies and Future Professional Life

Poster Abstract:
AI spreads into the learning environment of students. Internet providers integrate AI features that automatically summarizing content. These tools turn into filters hiding away the variety of sources and reducing the trustworthiness of generated content due to the statistical nature of AI and limited accessibility to reliable sources. Global Japanese Studies widen the knowledge on Japan while including critical thinking. This refers to searching, factchecking and applying standards for assessing information on Japan. Long and Magerko (2020) define AI literacy as “a set of competencies that enables individuals to critically evaluate AI technologies; communicate and collaborate effectively with AI; and use AI as a tool online, at home, and in the workplace.” This poster presents findings from a survey among first graders of the Global Japanese Studies Department regarding their use, expectations and concerns towards AI tools for their further study (Apr to Dec 2025, Japanese N=59 and foreign students N=50).

Biography:
Dr Susanne Brucksch is associate professor at the Department of Global Japanese Studies, Teikyo University. Before, she worked as principal researcher at the German Institute for Japanese Studies (DIJ) and senior researcher at Freie Universität Berlin, and visiting scholar at Waseda University in 2016, and at the Max-Planck-Institute (MPI) for Innovation and Competition in Munich in Nov 2019. Besides, she has been serving as co-organizer of the STS-Technology-Section “Technology and Society in Japan and Beyond” since 2015. Her research and teaching addresses topics such as health, social welfare, technology and society in Japan.

Prof. Dr. Yuko Murakami (Rikkyo University), Prof. Dr. Yuko Murakami (HIroshima University), Prof. Dr. Takeo Tatsumi (Open University of Japan), Prof. Dr. Tomohiro Inagaki (Hiroshima University)

Poster Title:
Education in AI ethics: focusing on the elders in Japan

Poster Abstract:
The authors are developing educational content on AI ethics for non-professionals, focusing particularly on older adults and the challenge of an LLM ‘badness’ indicator and AI algorithm ‘running wild.’ It can not be denied that bad uses of AI cannot be eliminated, necessitating technical, social, and legal regulation, alongside education for both humans and machines. The project aims to enhance a proactive lifestyle for every individual adults. Disparities in AI experience can be overcome, the authors propose, via an autonomously adjusting AI curriculum that generates customized ethical problems based on learner interests. This requires a filtering system to screen out ‘inadequate examples’ using ‘badness’ metrics for LLMs and AI while respecting public decency and personal beliefs.

Biography:
Yuko Murakami is a Professor in the Graduate School of Artificial Intelligence and Science at Rikkyo University. She holds a Ph.D. in Philosophy from Indiana University. Her background includes focusing on the philosophical logic of actions and moral reasoning. Her current research spans several areas, including Logic, Philosophy and history of logic, Philosophy of artificial intelligence, and Philosophy of science. A key focus is the philosophical analysis of social issues concerning information technology, drawing on perspectives from the philosophy and history of science and technology. Her competence also extends to Philosophy of technology and ICT education.

Moe Nishisako
Master Student, Master’s program in health and environmental sciences, Fukuoka Women’s University

Poster Title:
Exploring the Potential of S2Vec for Predictive Policing

Poster Abstract:
The research aims to identify urban and environmental factors that influence the risk of crime occurrence. A major challenge in this field is the time required to collect and preprocess diverse urban data for specific research questions and statistical models. This study explores the potential of S2Vec to address this issue by automatically extracting geographic features from map data using AI. It compares the conventional feature engineering approach with an S2Vec-based approach to evaluate whether S2Vec can reduce data preparation efforts while maintaining predictive performance. The methodology and comparative results will be presented in a poster, highlighting the potential of S2Vec as an efficient tool for geographic crime prediction research.

Biography:
Moe Nishisako is a master student of Fukuoka Women’s University.
Her research interests are GIS, Embedding, Crime Prevention, and AI.

Yuki Yotsumoto
Graduate School of Science and Engineering, Doshisha University

Poster Title:
GENTOO: LLM-based Generative Error Correction on ASR for Conference Presentations Leveraging Proceeding Papers

Poster Abstract:
Speech recognition for conference presentations often suffers from degraded accuracy due to domain-specific terminology. To address this, we propose a novel LLM-based automatic speech recognition (ASR) error correction method that leverages proceedings papers as external knowledge. Unlike slides, proceedings papers provide dense, detailed text and figures, which are highly effective for correcting specialized context. We evaluated two approaches: “full-paper reference,” which inputs the entire paper text and figures into the prompt, and “relevant-section reference,” which utilizes Sentence-BERT to retrieve only the top-1 most relevant section. Evaluated on the ACL60/60 dataset using GPT-4.1 mini, both proposed methods outperformed conventional baselines (no external knowledge and slide-based reference). Notably, the full-paper reference method achieved the lowest Word Error Rate (WER), delivering an approximate 20% improvement over the uncorrected baseline. Our findings demonstrate that leveraging proceedings papers significantly enhances ASR correction performance for academic presentations.

Biography:
Yuki Yotsumoto is a Master’s student in the Graduate School of Science and Engineering at Doshisha University. His research interests include Natural Language Processing (NLP) and Automatic Speech Recognition (ASR). Notably, he spent one year studying at the University of Helsinki, Finland, as a recipient of the “Tobitate! Leap for Japan” scholarship, gaining international academic experience. He is scheduled to present two papers at the Forum on Information Technology (FIT) in September 2026: one on this ASR error correction method, and another extending the same framework to machine translation. He aims to enhance multilingual communication in academic and professional settings.

Jibaek Lim^1,2,3,†, Joohyun Kim^4,†, Hyunsuk Chung^5,†, Seungyeon Ji^1,6, Sangeun Lee¹, Chaehee Kim¹, Yuna Koo⁷, Tae Youn Kim⁷, Chung Man Kim⁷, Soyeon Caren Han^5,*, Kyungreem Han^1,8,†,*, Sangwook Yi^2,3,9,*
¹Brain Science Institute, Korea Institute of Science and Technology, Seoul, Republic of Korea
²Department of Philosophy, Hanyang University, Seoul, Republic of Korea
³HY Center of Ethics, Law, and Policies for Science and Technology, Hanyang University, Seoul,
Republic of Korea
⁴Institute for Gender and Law, Ewha Womans University School of Law, Seoul, Republic of Korea
⁵School of Computing and Information Systems, University of Melbourne, Melbourne, Australia
⁶Department of Computer Science and Engineering, Korea University, Seoul, Republic of Korea
⁷Law Firm ELIM, Seoul, Republic of Korea
⁸University of Science and Technology KIST School, Seoul, Republic of Korea
⁹Department of Artificial Intelligence, Hanyang University, Seoul, Republic of Korea

^†These authors have contributed equally.
^*Corresponding authors

Poster Title:
Persona-based AI agent for bias mitigation in LVLMs

Poster Abstract:
Mitigating bias in AI agents based on Large Vision-Language Models (LVLMs) is essential to ensuring trustworthy and safe human–AI interactions. In this study, we propose a novel human-in-the-loop framework for bias-aware fine-tuning of AI agents. The framework comprises two key components: (1) a persona representation unit that explicitly models the AI agent’s intended role, functional objectives, legal responsibilities, and ethical principles, and (2) a self-evolving unit that continuously refines the AI agent to reflect current societal consensus. The framework processes multimodal inputs via an LVLM-based agent with multimodal encoders and conversational memory for context-aware reasoning. To support trustworthy decision-making, a bias governance module integrates bias ontologies, legal and ethical knowledge, and expert-defined taxonomies, providing structured guidance for bias-aware reasoning. This knowledge is incorporated into a knowledge-guided, parameter-efficient fine-tuning (PEFT) strategy that efficiently adapts pretrained LVLMs while preserving their general capabilities. Finally, a human-in-the-loop reinforcement learning mechanism continuously refines the agent using (explicit) expert feedback and (implicit) user interactions, enabling progressive alignment with evolving societal and regulatory standards. This enables the AI agent to identify, assess, and mitigate potential discriminatory outcomes based on protected grounds under anti-discrimination law before they cause real-world harm, thereby supporting the development of safe and trustworthy AI systems.

Biography:
Prof. Dr. Sang Wook Yi
Professor Sang Wook Yi studied physics for his B.Sc and M.Sc. at Seoul National University. He finished his PhD in philosophy of science at LSE, University of London. He is now a tenured professor at the department of philosophy and at the department of artificial intelligence, Hanyang University, while leading HY Center
for Ethics, Law and Policy of Science and Technology. Professor Yi’s research interests cover a wide range of topics in the philosophy of science and technology. He is currently working on various ethical issues relating to frontier research especially AI and synthetic biology.

Prof. Dr. Kyungreem Han
Kyungreem Han earned his BSc and PhD degrees from Seoul National University. Before joining the Brain Science Institute/Korea Institute of Science and Technology as principal investigator and professor (KIST School), he gained experience in theoretical physics at the Center for Theoretical Physics at SNU and in computational chemistry at the National Institutes of Health (MD, USA). His primary contributions include: i) physics-based interpretations of AI, ii) combining AI with multidisciplinary knowledge, and iii) AI safety and trustworthiness. His multidisciplinary team employs various theoretical and computational techniques, along with high-performance classical/quantum computing, to study information-processing principles in life and artificial life.

Prof. Dr. Caren Soyeon Han
Caren Han is an Associate Professor at the School of Computing and Information Systems, The University of Melbourne, and an Adjunct Professor at POSTECH and University of Edinburgh. She received her PhD in Computer Science, where she graduated with First Class Honours and the Chancellor’s Award. Before joining the University of Melbourne, she was an Assistant Professor at the University of Sydney. Her research focuses on natural language processing, multimodal foundation models, trustworthy AI, and AI agents. She leads international research collaborations with major technology companies and government organizations, including Google, Microsoft, Amazon, NASA. Her work aims to develop reliable, interpretable, and human-centered AI systems that integrate advances in multimodal learning, reasoning, and real-world decision making.

Prof. Dr. Joohyun Kim
Joohyun Kim is a Research Professor at the Institute for Gender and Law, Ewha Womans University School of Law. She earned her B.A., M.A., and Ph.D. in Law from Ewha Womans University. Her scholarship spans philosophy of law, human rights law, and gender law, with a focus on equality and anti-discrimination law. Her recent research has centered on artificial intelligence, including an LLM-based experimental study on the gender sensitivity of AI judges, a critique of South Korea’s AI Framework Act through a data feminism lens, and an examination of regulatory strategies for mitigating gender bias in AI voice assistants.

Jibaek Lim
Jibaek Lim is a master’s student at the Department of Philosophy, Hanyang University in Korea and a researcher at the Brain Science Institute/KIST. He received his bachelor’s degree in Computer Science with a minor in Physics from Hanyang University. His research interests include Artificial Intelligence Humanities and Artificial Intelligence Art.

Dr. Tamon Nakano
Research Staff, The Institute for AI Safety and Security, DLR German Aerospace Center

Poster Title:
Reservoir Computing: Safe, Private, Energy-Efficient Edge AI by Design

Poster Abstract:
Reservoir Computing is a lightweight kind of AI. Because it is so light, it can run directly on small, everyday devices, rather than the large, power-hungry computers and cloud servers mainstream AI relies on. That is what makes it safe, private, and energy-efficient at once.
Energy-efficient and green: it runs on an ordinary processor, or even a tiny chip, using a fraction of the power large AI needs.
Private: because it runs inside the device itself, such as medical or customer equipment, the data never has to be sent to the cloud.
Safe: its result is a clear formula you can read and check, not a black box, so you can see why it gives each answer.
On the poster, we show examples: estimating hard-to-measure values inside machinery, spotting early signs of equipment failure, predicting the remaining life of engines and batteries, and reading blood pressure on a wearable.

Biography:
Tamon Nakano is a researcher working on reservoir computing for industrial and edge AI, affiliated with DLR (the German Aerospace Center). He has built his career across Japan, France, and Germany, giving him a trilateral perspective that fits this symposium. He is also co-founder of Entrox Systems, a DLR spin-out startup applying this approach in industry through collaborations spanning the three countries. His interest lies in lightweight, easy-to-understand AI that runs on small devices without sending data to the cloud, and in advancing reservoir computing worldwide, with Japan, France, and Germany at its center.

Maida Aizaz
Master’s Student, Graduate School of Data Science, KAIST

Quang Minh Nguyen
Master’s Student, Graduate School of Data Science, KAIST

Braahmi Padmakumar
Master’s Student, School of Computing, KAIST

Poster Title:
The Broken Telephone Changes Tone: Examining Nuanced Linguistic Cues in LLM Chains-of-Translation

Poster Abstract:
As LLM-generated content proliferates online, texts are increasingly subject to repeated processing and translation by models, necessitating the understading of how this reshapes language. Prior work has revealed degraded factual content and reduced diversity, but the fine-grained linguistic shifts underlying these effects remain unexplored. We track changes in epistemic markers, grammatical voice, degree adverbs, and nominalisation density across 12 iterations of round-trip translation applied to 600 BBC News articles, varying intermediate language, translation model, and chain topology across 17 configurations. We find an epistemic shift: evidential and factive markers increase while hedges decline, potentially causing tentative claims to read as more certain. Texts undergo register-level formalisation: informal degree adverbs give way to formal alternatives, active-voice density drops, by-phrase passives attrite disproportionately, and nominalisation density rises—alongside clear model-specific patterns for certain settings. These shifts erode source, register, and agency markers, offering a fine-grained account of the factual degradation reported previously.

Biography:
Maida Aizaz
Her name is Maida, and she is an M.S. Data Science student at KAIST, advised by Professor Lanu Kim, with a background in Industrial Design and AI. Her research sits at the intersection of computational social science, NLP, and fairness—exploring topics like geographical bias in academia, LLM persona generation, geopolitical framing in academia, and forensic facial reconstruction. She has published and presented at venues including ICML, ACL, CVPR, and EMNLP. Her goal is to blend design thinking and data-driven methodologies to uncovering systemic disparities in science and society.

Quang Minh Nguyen
Minh is a master’s student at KAIST Graduate School of Data Science. His research is on AI safety and interpretability, with published work at ACL, EACL, and COLM. He is currently interested in the introspective capabilities and monitorability of language models.

Braahmi Padmakumar
Braahmi is a Master’s student at the School of Computing in KAIST, advised by Professor Joseph Seering. With a multidisciplinary background in Computer Science, Industrial Design, and Science & Technology Policy, her research explores the intersection of technology and lived experience, with a focus on how emerging technologies can be built to serve human wellbeing. She is currently interested in trust and safety in digital spaces, content moderation, and building persona-grounded simulations and multi-agent systems.

Prof. Dr. Philippe Greciano
President, Franco-German University

Poster Title:
The Franco-German University : a strategic and long-term partner with Japan

Poster Abstract:
The poster will present the Franco-German University (FGU) which is developing numerous AI projects in Europe and around the world (conferences, round table discussions, expert meetings). It is a privileged partner for organizing Franco-German and Japanese events in this field of excellence (e.g. 2024 French-German and Japanese AI Symposium Tokyo ; 2026 French-German and Japanese Workshop, Tohoku University).

Through its range of programs and scientific events, the FGU is training many AI specialists with the aim of overcoming current global challenges, including the protection of privacy protection, democracy, security and major social and environmental changes. Its network comprises more than 200 universities worldwide. It is developing several international activities with scientific and economic players and civil society.

Scientific Network in Japan : https://www.dfh-ufa.org/fr/news/exzellenz-partnerschaft-in-japan

Scientific Network in Europe : https://www.dfh-ufa.org/fr/communiquesdepresse/dialogue-franco-allemand-sur-lintelligence-artificielle-luniversite-franco-allemande-au-coeur-de-lactualite

Biography:
Prof. Dr. Philippe Gréciano is Full professor, Jean Monnet Chair of Excellence. He is President of the Franco-German University (FGU) and strongly committed to international scientific and economic cooperation. He develops cooperations with Japan and offers opportunities for professors and researchers to work together from a Franco-German, International and European perspective.

Nicolas Fliegel
Special Research Student (visiting graduate researcher), Graduate School of Information Science and Technology – Department of Mechano-Informatics (Takeuchi Laboratory), The University of Tokyo

Poster Title:
The Right to be Forgotten: Evaluating the Effectiveness of Federated Unlearning Algorithms

Poster Abstract:
Federated learning enables collaborative model training without sharing raw data, supporting GDPR compliance. Yet enforcing the “Right to be Forgotten” (Art. 17 GDPR) is hard: removing a client’s data influence from a trained model without full retraining often leaves recoverable traces. This work asks what actually determines federated unlearning effectiveness. I built a benchmarking framework and ran controlled experiments on a representative unlearning algorithm (subspace-based federated unlearning), using membership inference attacks as a privacy audit. Across a grid of system parameters (data heterogeneity, client participation) and algorithm hyperparameters, system conditions dominate: under balanced data and full participation, attack success falls to ~50% (random guessing) with 98.9% accuracy retained, matching full retraining; under skewed data and low participation, traces persist (60.5% attack success, 7.3% accuracy loss). Federated unlearning is therefore a systems problem. The framework and guidelines support practitioners deploying privacy-compliant federated systems.

Biography:
Nicolas Fliegel is a computer science master’s student at the Technical University of Munich (TUM) and an incoming Special Research Student at the University of Tokyo’s Graduate School of Information Science and Technology (Takeuchi Lab, biohybrid systems) from October 2026. He completed the federated unlearning research presented here during his bachelor’s studies at the Karlsruhe Institute of Technology (KIT), at the KASTEL Institute of Information Security and Dependability. His research interests span privacy-preserving and trustworthy machine learning, federated systems, and embodied AI. He also works on quantum technology and society at TUM’s Quantum Social Lab.

Dr. Kenji Imamura
Senior Researcher, Universal Communication Research Institute, National Institute of Information and Communications Technology

Poster Title:
Toward Multilingual and Multicultural Evaluation for AI for Safety and AI Localization

Poster Abstract:
AI performance can vary across languages and cultures. To address this issue, we are developing two evaluation datasets for assessing AI for safety and AI localization. The first dataset consists of disaster preparedness quizzes. Since incorrect responses in disaster preparedness can have serious consequences, performance on such quizzes is an important indicator of AI for safety. We constructed this dataset in a multiple-choice format with four answer options. The second dataset is a culture-aware benchmark. In collaboration with ASEAN countries, we plan to develop a multimodal quiz benchmark in which the correct answers may differ depending on the region or cultural context.

Biography:
Kenji Imamura received his Ph.D. in Engineering from the Nara Institute of Science and Technology (NAIST). He is currently a Senior Researcher at the National Institute of Information and Communications Technology (NICT). His research interests include natural language processing, particularly machine translation.

Viet Qui Le
Graduate Student, Institue for Advanced Study of Human Biology (ASHBi), Kyoto University

Poster Title:
Towards Highthroughput AI-based analysis methods for characterisation of embryo models.

Poster Abstract:
Pluripotent stem cell-based in vitro models enable unprecedented opportunities to study early human development while addressing ethical and technical constraints. But, characterizing these models quantitatively remains challenging. Current image-based quantitative assessments e.g. morphometric measurements, remain largely manual, subjective and difficult to scale. With rising high-throughput imaging data, automated, reproducible pipelines become a necessity. While AI-based methods show potential their implementation requires specialized technical expertise and curated datasets. To address this, we are developing an ongoing AI-driven image analysis pipeline for the morphometric characterization of in vitro models. Our preliminary framework streamlines image filtering, classification, and quantification, while integrating tools for bulk annotation and dataset curation to support research workflows. The goal is to facilitate time-efficient analysis, for deeper insights into early development. Future iterations are planned to include finer anatomical segmentation and support for time-resolved data to ultimately, promote accessible AI adoption and high-quality datasets for next-generation biomedical imaging.

Biography:
After obtaining a Bachelor of Sciences focused on Biology at Maastricht University, he is now pursuing a Master’s in Medical Science at the Alev Lab in Kyoto University’s ASHBi. His research focuses on the morphological characterization of embryo models. To improve data quality and efficiency, he developed AI-based pipelines that automate image analysis, replacing tedious and time consuming methods. With AI becoming more vital to our work, his job is implementing these tools in the lab, to deepen our understanding of early human development. Aiming to apply this knowledge for e.g. modelling diseases or improving synthetic organ development.

Prof. Dr. Se-Young Yun
Associate Professor, Graduate School of AI, KAIST

Poster Title:
UniSAFE: A Comprehensive Benchmark for Safety Evaluation of Unified Multimodal Models

Poster Abstract:
Unified Multimodal Models (UMMs) offer powerful crossmodality capabilities but introduce new safety risks not observed in single-task models. Despite their emergence, existing safety benchmarks remain fragmented across tasks and modalities, limiting the comprehensive evaluation of complex system-level vulnerabilities. To address this gap, we introduce UniSAFE, the first comprehensive benchmark for system-level safety evaluation of UMMs across 7 I/O modality combinations, spanning conventional tasks and novel multimodal-context image generation settings. UniSAFE is built with a shared-target design that projects common risk scenarios across task-specific I/O configurations, enabling controlled cross-task comparisons of safety failures. Comprising 6,802 curated instances, we use UniSAFE to evaluate 15 state-of-the-art UMMs, both proprietary and open-source. Our results reveal critical vulnerabilities across current UMMs, including elevated safety violations in multi-image composition and multi-turn settings, with image-output tasks consistently more vulnerable than textoutput tasks. These findings highlight the need for stronger system-level safety alignment for UMMs.

Biography:
Se-Young Yun is an Associate Professor at the Kim Jaechul Graduate School of AI, KAIST, where he has been a faculty member since 2017. Prior to joining KAIST, he was a postdoctoral researcher at Los Alamos National Laboratory, the MSR-INRIA Joint Centre, and KTH Royal Institute of Technology. He has also been a visiting researcher at the London School of Economics and Political Science (LSE) and Microsoft Research (MSR). His research interests include efficient large language models, AI for science, and AI for mathematics.

Dr. Judit Erika Magyar
Country Representative, International, EURAXESS Japan

Poster Title:
Unlocking EU-Japan AI Collaboration: Horizon Europe Cluster 4 and EURAXESS Japan Services

Poster Abstract:
With Japan’s historic 2026 association to Horizon Europe, Japanese institutions can now participate as equal partners in Pillar II collaborative research. This poster maps out multi-billion euro funding tracks for Artificial Intelligence, data, and robotics under Cluster 4 (Digital, Industry, and Space) for the 2026–2027 Work Programme. It highlights key avenues like GenAI4EU and the RAISE pilot for AI in science. Navigating these massive European frameworks requires local support. To fill this gap, this presentation showcases how EURAXESS Japan acts as a free, specialized gateway for researchers, demonstrating how these services—including matchmaking portals, partner searches, bi-monthly flash notes, and dedicated proposal workshops—empower Japanese innovators to build multinational consortia. Ultimately, this poster provides a practical roadmap for Japanese universities, startups, and enterprises to secure direct EU funding and lead global AI breakthroughs.

Biography:
Judit Erika Magyar has been the Country Representative of EURAXESS Japan for 7 years, promoting the European Union as an excellent destination for early, mid- and advanced career researchers of any discipline.
Prior to her current post, she worked in project management and lectured extensively at universities in Europe and Japan. She completed her various MA degrees in Pecs, Budapest, Milan and Fukuoka before pursuing PhD studies in Tokyo with a focus on Law, International Relations and History, respectively. Having spent 20 years in Japan, she has a thorough knowledge about the local academic, research and innovation scenes. Since June 2025 she has also been serving as Regional Liaison Officer at MSCA-GLOPOL.