Global AI Scholars and Samsung's latest major presentations
Research activities
Samsung Electronics today announced the opening of the Samsung AI Forum 2025, which will be held from September 15th to 16th. In its current ninth year, the forum will serve as a global venue for leading scholars and industry experts to bring together AI's latest breakthroughs and explore future research directions.
Young Hyun Jun, Vice Chairman and CEO of Samsung Electronics, said, “Samsung is applying AI throughout our business to develop the fundamental technologies that make AI more intuitive and seamless. “This year's Samsung AI Forum brings together key industry and academia experts to discuss how AI is transforming society and industry and share insights about what we expect will be a meaningful exchange of ideas.”
This year's forum features keynote speeches from world-renowned AI scholars. Includes Yoshua Bengio, a professor at the University of Montreal and pioneer of deep learning. Like Stefanoelmon, a professor at Stanford University and co-founder of Startup Inception, he led the development of a diffusion-based language model (DLM).
Day 1: Global Scholars explore the future of AI semiconductors
Samsung's Device Solutions (DS) Division held the first day of the forum under the theme of “Vertical AI Strategy and Vision for the Semiconductor Industry” near Samsung's semiconductor site in Yongin, South Korea.
Professor Bengio gave a keynote speech outlining the broader risks of today's AI models, including the ability to bypass human control and the possibility of misuse. As a parent, he introduced Scientist AI, a new model designed to alleviate such concerns.
“Unlike human imitation and willingly constructed models, scientist AI focuses on providing truthful answers based on verified facts and data,” Professor Bengio said, highlighting the potential of models that increase AI safety and accelerate scientific discovery. He also shared the latest findings on AI on material development in keynote speech.
Amit Gupta, Senior Vice President of Siemens EDA, led a session entitled “The Future of AI-Driven Electronic Design.” During his presentation, he highlighted the importance of integrating AI into Electronic Design Automation (EDA) tools, noting that end-to-end systems across the entire workflow are key to reaching the full potential of AI.
The technical sessions continued, featuring Executive Vice President and Head of the AI Center for DS, Professor Seokhyung Kang of Pohan University of Science and Technology (Postech), and Professor Il-Chul Moon of Korea Advanced Institute of Science and Technology (KAIST). Each speaker shared recent research and development in AI applications for semiconductor design and manufacturing, while also providing a perspective on field evolution.
“AI is already an essential tool for chip design and software development,” EVP Song said. “As semiconductor manufacturing becomes more complex, we expect it will help AI to address the technical challenges that arise.”
The Forum also recognized three recipients of the Samsung AI Researcher of the Year Award. Professor Nicholas Papanott at the University of Toronto, Professor Rose Yu at the University of California, San Diego, and Professor Lerrell Pinto at New York University. All award winners were invited to present their work and share insights from the latest research.
Day 2: Focusing on the age of agent AI to increase productivity
Day 2 of the forum will be held online by Samsung's Device Experience (DX) division and will be streamed live on the Samsung developers' YouTube channel under the theme of “generated by Agent AI.”1
“Generating AI has already become an essential tool across daily life and industries,” said Paul (Kyungwhoon), Cheun, CTO of Samsung Electronics' DX division and Head of Samsung Research. “As we enter the age of agent AI, Samsung will continue to focus on developing AI technologies that offer tangible benefits to its users.”
Day 2 will include keynote speeches from Professor Joseph E. Gonzalez at UC Berkeley, as well as leading researchers in language modeling and AI agents. Subbarao Kambhampati, a professor at Arizona State University and a global authorities for AI planning and decision-making. Stefano Elmon, a professor at Stanford University and co-founder of Inception.
Professor Gonzalez presents research into enhancing agent capabilities in large-scale language modeling (LLM)-based systems. In particular, he introduces a paradigm of sleep time calculations. This allows agents to use idle interaction intervals to infer, learn and plan.
Professor Kambhampati shares his research on large-scale inference models (LRMS) aimed at addressing the limitations of LLMS. He points out that while current models are excellent at text generation, they remain limited in actual accuracy, planning and complex inference. It highlights key challenges such as guaranteeing answer guarantees, enabling calculation of context adaptation, and providing interpretation of intermediate inference procedures.
Professor Elmon introduces a diffusion language model that applies the widely used diffusion models for image, video and audio generation to the language domain. This approach aims to overcome the limitations of traditional sequential text generation methods and propose a more efficient paradigm of language models.
During the technical session, Samsung Investigators will present the latest developments, including:
- Camera AI technology for automatic color temperature control
- Knowledge distillation-based methods for more efficient training of LLM and its applications
- Device AI technology designed to guide LLMS to home appliances such as smartphones and televisions
- Auto-dubbing technology that generates narration with original speaker audio
- Deep Dive technology that uses multi-agent systems to analyze and automatically generate various reports
- Document AI technology that automatically converts various document formats into structural data for LLMS and agent systems
- Device AI Studio for Developers to Shorten Development Cycles for Generated AI Models
1 An intelligent system that allows you to make autonomous decisions and perform tasks yourself.
