"Supermassive AI for a Better Tomorrow"... 7th 'Samsung AI Forum'

Experts in Artificial Intelligence and Computer Science Attend

In the era of ultra-large artificial intelligence (AI), a forum was held to discuss the future direction of next-generation semiconductors and the research achievements so far.

Samsung Electronics CEO Kyung Kye-hyun delivers the opening address at the 'Samsung AI Forum' held at the Suwon Convention Center in Yeongtong-gu, Suwon-si, Gyeonggi-do on the morning of the 7th.
[Image source=Yonhap News]

On the 7th, Samsung Electronics held the 7th 'Samsung AI Forum 2023' at the Suwon Convention Center in Suwon, Gyeonggi Province. The Samsung AI Forum is a platform where world-renowned scholars and experts in AI and computer engineering (CE) gather to share the latest research trends and explore future innovation strategies. This year, over 1,000 people, including students, attended the event.

At the opening remarks, Kyung Kye-hyun, Vice Chairman of Samsung Electronics, said, "Generative AI technology is rapidly emerging as an innovative means to solve various problems faced by humanity," adding, "There is a growing need for in-depth research on the safety, trustworthiness, and sustainability of the technology."

Keynote Speeches by Professor Yoshua Bengio and CEO Jim Keller

This year’s forum theme was 'Ultra-large AI for a Better Tomorrow.' Ultra-large AI refers to large-scale AI systems characterized by processing vast amounts of data and tasks. The semiconductor industry expects that due to the exponential increase in computing throughput caused by ultra-large AI, next-generation semiconductors will require unprecedented levels of high performance, large capacity, and low power consumption.

Speakers and attendees discussed AI and CE technology research achievements and future development directions to strengthen next-generation semiconductor research capabilities under these challenges. Professor Yoshua Bengio of the University of Montreal, Canada, a global AI scholar, and Jim Keller, CEO of Tenstorrent, a Canadian AI semiconductor startup, delivered keynote speeches.

Professor Yoshua Bengio gave a lecture titled 'Towards Safe AI Researcher Systems,' introducing safe AI machine learning algorithms that prevent AI technology outcomes, evolving based on large language models (LLM), from diverging from the developers’ intended research goals.

CEO Jim Keller presented ways to overcome AI technology limitations through next-generation semiconductor design innovation under the theme 'Own Your Silicon.' He emphasized new possibilities for next-generation AI through innovative hardware architecture design based on open hardware design assets (RISC-V).

Samsung Electronics SAIT Envisions AI Utilization and Future Changes in Semiconductors

Samsung Electronics’ fundamental technology research organization, SAIT (formerly the Comprehensive Technology Institute), held detailed sessions on AI and CE under the themes 'Changes in LLM and Industrial AI' and 'Ultra-large Computing for LLM and Simulation.' These sessions featured lectures not only from global scholars and experts but also from research leaders of Samsung Electronics SAIT AI Research Center and System Research Center.

Samsung Electronics SAIT research leaders outlined plans for AI utilization, including LLM, across semiconductor development in the AI field and envisioned future changes in semiconductors. In the CE field, they discussed the potential for future computing advancements such as process simulation using AI.

On the same day, Samsung Electronics SAIT announced the recipients of the 'Samsung AI Researcher Award,' which discovers outstanding global young researchers in AI, and the winners of the 'Samsung AI·CE Challenge,' conducted to foster domestic AI talent.

Five individuals, including Professor Jason Lee of Princeton University, USA, were selected for the Samsung AI Researcher Award. Professor Jason Lee focuses on theoretical and applied research in AI fields such as deep learning, reinforcement learning, and optimization. He has contributed significantly to the advancement of global AI research by publishing numerous excellent papers in these fields. In the Samsung AI·CE Challenge, which saw participation from 410 teams (1,481 students), a total of 16 teams were awarded.