This research group focuses on the advancement of reinforcement learning systems, with efforts concentrated on discovering the field of Logical Reinforcement Learning (LRL) and enhancing reinforcement learning schemes. In reinforcement learning, the goal is to improve algorithm efficiency, allowing systems to effectively handle complex sensory inputs such as images and audio. In parallel, the group’s work in logical reinforcement learning aims to provide novel techniques for green and democratic decision-making algorithms, using the benefits of traditional computation chips specialization in logical operations.
A significant aspect of the research involves the combination of the Tsetlin Machine and reinforcement learning techniques. The Tsetlin Machine, known for its simplicity and use of propositional logic, offers clear interpretability and reduced computational demands. By incorporating this machine into reinforcement learning strategies, the group aims to develop transparent systems that are adept at learning from their environment, marking a step forward in the quest for machines that exhibit a level of intelligence and understanding comparable to human reasoning, all within a framework of green and democratic AI.
cair to join us, please contact:
Associate professor Per-Arne Andersen: per.andersen@uia.no
