Dr Diego Perez Liebera

What is it about Statistical Planning and Reinforcement Learning that you enjoy so much? 

What I particularly enjoy about teaching Statistical Planning and Reinforcement Learning is that it closely aligns with my own research. I work on statistical forward planning methods in games, so I can bring cutting-edge research questions, methods, and insights directly into the classroom. This creates a very natural link between research and teaching, and it allows students to see how the material they are learning is actively being developed and applied.

I also find that games provide a uniquely engaging environment for learning AI. They offer clear objectives, measurable outcomes, and controlled complexity, which make them ideal testbeds for experimentation. Students can see algorithms in action, observe how agents learn and plan, and immediately connect theory to behaviour. This makes the learning process both rigorous and enjoyable, fostering deeper understanding while maintaining a strong sense of curiosity and motivation.

Why do you think this module is critical for students to learn? What will they gain from it? 

I believe this module is critical because it introduces students to a general form of Artificial Intelligence that is applicable across a wide range of problems and domains. Rather than focusing narrowly on a single technique, it equips students with conceptual tools that can be transferred to many real-world challenges. Beyond the technical content, students also gain core scientific skills: how to formulate research questions, design and conduct experiments, analyse results rigorously, and draw sound conclusions. These are foundational abilities not only for AI, but for any evidence-based discipline.

How would you describe your teaching style? 

I tend to teach in the way I would like to be taught myself: through clear examples, well-designed diagrams, and a strong focus on the key ideas that truly matter. When a concept is particularly important, I make a point of explaining it from more than one perspective, so that students with different ways of thinking can understand it. I also ensure that theoretical material is consistently reinforced through practice. In the labs, students implement and experiment with the ideas we discuss in lectures, which helps consolidate understanding and builds confidence in applying theory to concrete problems.

What components of the module would be transferable to the industry students end up in? 

Several components of the module translate directly to industry contexts. Planning algorithms, for example, underpin decision-making systems in logistics, robotics, and software systems more broadly. The ability to reason under uncertainty, dealing with hidden information and stochastic environments, is equally vital. Many real-world systems are incomplete, noisy, or probabilistic, and learning how to model and reason within such settings prepares students for challenges they will encounter in technology, finance, healthcare, and beyond.

What advice do you have for students considering a degree in Artificial Intelligence?

Students considering a degree in Artificial Intelligence should be prepared to engage seriously with programming, and have a solid base on mathematics. AI is a practical discipline, but it is also grounded in formal reasoning. Developing confidence in coding and in mathematical thinking will make the journey far more rewarding. I would also encourage curiosity. Read widely, go beyond the core lecture notes, and explore supplementary materials. AI evolves rapidly, and cultivating the habit of independent learning is essential for long-term success in the field.

What has been the highlight of your time at EECS? 

One of the highlights of my time at EECS has been the opportunity to design my own modules. This allows me to bring the latest research directly into the classroom and to update the content each year so that it remains current and relevant. Research-wise, I have particularly valued the opportunity to collaborate with companies and industry partners. Working at the intersection of Games AI research and real-world applications has been immensely rewarding, as it demonstrates how ideas developed in games can transfer to broader societal and industrial challenges.