Keynote Talks
Keynote Talks
Prof. Liu Zhiyuan
Tsinghua University, China
Prof. Liu Zhiyuan
Liu Zhiyuan is a Tenured Associate Professor in the Department of Computer Science at Tsinghua University. His main research interests cover large language models, knowledge graphs, and social computing. He has published over 200 papers in international journals and conferences on AI with over 67k citations and has released series of MiniCPM and MiniCPM-o/v models specifically designed on device. He was listed as an Elsevier China Highly Cited Scholar for five consecutive years (2020–2024) and in MIT Technology Review's 35 Innovators Under 35 (China).
Talk Title: Towards Artificial General Intelligence: Key Challenges and Trends
Talk Abstract: In early 2025, DeepSeek V3 and R1 released by China’s DeepSeek Team attracted global attention. This demonstrates that Chinese teams can maximize the utilization of limited resources through algorithmic innovations despite constraints from computing power bottlenecks, achieving advanced model capabilities with low cost and high quality. These two models reflect two main themes in the development of artificial intelligence: enhancing model intelligence through algorithmic innovations to expand the capability spectrum of AI; achieving continuous improvement in the capability density of models, rapid reduction in training and application costs, and high-quality development of AI through synergistic innovation across data, models, and computing power—laying the foundation for AI inclusivity. This talk will explore the development trends of large models moving towards artificial general intelligence from these two perspectives.
Prof. Shin'ichi Satoh
National Institute of Informatics (NII), Tokyo, Japan
Prof. Shin'ichi Satoh
Shin'ichi Satoh is a professor at National Institute of Informatics (NII), Tokyo. He received PhD degree in 1992 at the University of Tokyo. His research interests include image processing, video content analysis and multimedia database. Currently he is leading the video processing project at NII, addressing video analysis, indexing, retrieval, and mining for broadcasted video archives
Talk Title: Psychiatric Disease Diagnosis: The Challenge of Multimedia Technologies
Talk Abstract: Psychiatric diseases are a group of important diseases that reduce the quality of life of human beings due to their high incidence and long duration. Although it is hoped that they will be overcome as soon as possible, unfortunately, the current situation is far from that. One of the reasons for this is the lack of biomarkers suitable for diagnosing and assessing the severity of mental illness. Diagnosis and severity assessment are performed through conversations between psychiatrists and patients, and are heavily dependent on the doctor's experience. Such methods lack objectivity and quantification, leading to various problems such as inconsistent diagnoses and unclear criteria for starting treatment. In response to these problems, we have been working on the application of multimedia technologies including natural language processing (NLP), computer vision, lifelog, for psychiatric disease detection. We have been conducting multifaceted studies, such as preparing benchmark data for automatic diagnosis of mental illness, examining automatic diagnosis technology using NLP and image analysis, analyzing the mental state of society through SNS analysis, and analyzing mental and physical health conditions observed from dietary content. The talk will showcase couple of our approaches.
Prof. Cathal Gurrin
Dublin City University, Ireland
Prof. Cathal Gurrin
Professor Cathal Gurrin is an academic at Dublin City University (DCU) in Ireland and the Deputy Director of Ireland's ADAPT Centre for digital content technologies. Gurrin’s research interests focus primarily on personal analytics and lifelogging, which involves creating extensive personal databases of lifelog images and other sensor data to capture and analyse daily activities and experiences. His work aims to develop assistive technologies that use wearable sensors and data analytics to infer knowledge about real-world activities and enhance individual performance, health and safety. Gurrin is heavily involved in conference and challenge organising and he will be the general co-chair of both ACM MM'25 and ACM Web'27 in Dublin. He is also a organiser of the annual VBS Challenge at MMM and co-founder of the ACM LSC Challenge at ICMR.
Talk Title: Learnings From a Decade of Research into Lifelog Organisation and Retrieval
Talk Abstract: In recent years, the organisation and retrieval of lifelog data has emerged as a new challenge and opportunity within the multimedia research community. A lifelog is a comprehensive digital record of an individual’s daily activities, captured through wearable sensors, mobile devices, and software logs, encompassing images, location, biometrics, and more. In this talk I will introduce the challenges of organising and supporting retrieval from multimodal lifelogs, while also highlighting many of the promising use-cases of the technology. I will explore the progress made as a result of various international challenges, such as the ACM Lifelog Search Challenge (LSC) and highlight the evolution of retrieval techniques, system architectures, user interfaces, and evaluation methodologies developed by the community. By examining the insights gained from multiple years of progress, this talk will offer a comprehensive overview of the state-of-the-art in lifelog retrieval and discuss the implications for future research at the intersection of multimedia, human memory augmentation, and AI-driven personal data organisation. We will take a forward-looking perspective on the opportunities and open questions that lie ahead within this emerging field of research.
© 2025 ACM Multimedia Asia Conference. All Rights Reserved.