Reinforcement Learning: Shaping the future of Gaming and Robotics

The NDLI Club of DBCE in collaboration with the Department of Electronics and Computer Science conducted a session on “Reinforcement learning: Shaping the future of Gaming and Robotics” on 30^th August 2024. The session was delivered by Ms. Mathilda Colaco, Assistant Professor in Electronics and Computer Science. The NDLI session was organised by Ms. Trima P. Fernandes e Fizardo, Assistant Professor in Electronics and Computer Science.

The presentation focused on the applications of this cutting-edge technology, emphasizing its transformative impact on these fields. The session commenced by introducing the core concepts of reinforcement learning, such as action, state, reward, agent, and environment. The speaker illustrated these concepts through engaging examples like the “Mouse and Cheese” game and the “Frozen Lake” game, helping attendees grasp the basic mathematics involved in reward optimization.

The session also delved into the balance between exploration and exploitation, exploring new possibilities and utilizing known strategies, using relatable analogies like choosing dishes at a restaurant to explain decision-making in reinforcement learning. The presentation highlighted real-world applications, including the AlphaGo AI that famously defeated a world champion Go player, demonstrating the power of reinforcement learning in complex decision-making scenarios.

In the context of robotics, the applications of reinforcement learning in tasks like robotic locomotion and autonomous driving was discussed, showcasing examples like an ant robot learning to walk and self-driving cars, illustrating the potential of this technology to revolutionize robotics.

The session was both informative and engaging, providing attendees with a comprehensive understanding of reinforcement learning and its future implications in gaming and robotics.