First row standing: Prof Randy Goebel, Founder of Openmind Research Institute (ninth from left), Prof Datuk Dr Ewe (tenth from left) Second row standing: Joseph Modayil, President and Founder of Openmind Research Institute (ninth from left) Prof Richard Sutton, Founder of Openmind Research Institute (screen)
An intensive programme on reinforcement learning, brought together 60 international participants including students, researchers and industry practitioners to the Sungai Long campus of Universiti Tunku Abdul Rahman (UTAR).
Called the “Openmind Winter School”, the four-day programme was UTAR’s first introduction to reinforcement learning.
A core subfield of artificial intelligence (AI), reinforcement learning is a machine-learning technique where a robot or AI programme learns to make decisions by interacting with an environment, using trial and error to maximise long-term rewards.
Examples of reinforcement learning use can be found in the display of online ads that are programmed to reach a certain target audience.
It is also used to study stock market patterns to develop better investment strategies.
The event was jointly organised by Canadian-based Openmind Research Institute (ORI), UTAR, AI Malaysia and Universiti Malaya (UM).
It was supported by AI Singapore, Universiti Teknologi Malaysia, National University of Singapore, Nanyang Technological University and Asean Academy of Engineering and Technology.
The Winter School was sponsored by Silverlake Axis, Berjaya Corporation, Foundation for Science and AI Research (SAIR) and Oriza Holdings.
The keynote address titled “Reinforcement Learning: AI from Experience and First Principles” was presented virtually by ORI co-founder Prof Richard Sutton, who also held a live “Ask Me Anything” session.
Regarded as the father of reinforcement learning, Sutton was recipient of the 2024 Association for Computing Machinery’s Turing Award for his contributions to its conceptual and algorithmic development.
During the programme, participants gained insights from ORI founders Randy Goebel and Joseph Modayil on the evolution of AI, advancing research and real-world applications.
UM’s Prof Chee Seng Chan was also on hand to showcase Malaysia’s first multimodal large language model (MLLM), which are algorithms that can understand and generate various forms of content ranging across text, images, video and audio.
Divided into two segments, the first saw lectures and hands-on lab sessions led by experts from ORI and University of Alberta.
The other part consisted of seminars by researchers from institutions such as Google and Eindhoven University of Technology.
The spectrum of topics ranged from core algorithms to the evolution of AI challenges, MLLM-based agents and the future of autonomous systems.
Participants also had the chance to use what they had learned when they implemented algorithms on a four-legged robot, at first via simulation and later, directly onto the hardware.
The tutorial, led by Dr Sorina Lupu from ORI, led to a lively robot competition.
In his closing address, UTAR president Prof Datuk Dr Ewe Hong Tat reflected on the programme’s success in fostering open-minded collaboration and scholarly exchange.
Prof Ewe expressed his heartfelt gratitude to the organisers, partners, sponsors and participants, emphasising his confidence that connections forged and ideas generated during the programme would continue to inspire and advance the AI community in the years ahead.
“UTAR celebrates the resounding success of the inaugural Openmind Winter School, which created a dynamic platform for knowledge exchange and professional networking across the AI research community.
“Plans are already underway for future editions, building on the programme’s momentum and impact,” he said.
