Alvaro Torralba Arias De Reyna

Project title

ConAn: Contrastive Analysis for State-Space Exploration

What is your project about?

Broadly speaking, my project focuses on advancing Artificial Intelligence for automated decision making. Specifically, my goal is to create general-purpose tools that can directly solve any given problem. And these tools must be safe to use: they must always guarantee that the decisions are correct and meet all the user's requirements. To achieve this, tools need to "think ahead", considering possible alternative decisions, predicting outcomes, and determining subsequent actions. 

This project will introduce a new concept to guide this thinking process: Contrastive Analysis. The key idea is to give AI tools the ability to understand what the advantages and disadvantages of each alternative are. My hypothesis is that this ability is a fundamental skill for decision making and could change the way current AI systems make decisions.

How did you become interested in your particular field of research?

I became interested in automated planning (the subfield of Artificial Intelligence in which I work) when I was finishing my undergraduate degree. What attracted me to the field was the idea of "general problem solving": Can we build tools that are intelligent enough to decide what is the best decision in any situation? 

It is often said that "automated planning is the art of thinking before acting", and the idea of creating programs that can "think" and solve a problem they have never seen before was very appealing to me. Imagine that, right after explaining the rules of chess to someone, they would be able to play perfectly; that’s what we are trying to achieve.

What are the scientific challenges and perspectives in your project?

There are, of course, many technical challenges that we must overcome to develop the new concept of Contrastive Analysis. On the theoretical side, we aim to define entirely new ways of comparing alternative decisions and understand their properties. On the empirical side, we will develop new algorithms based on this concept and evaluate how they can advance the state of the art in many applications.  

But in my opinion, the main obstacle is that we are breaking the rules: challenging some of the assumptions behind well-established algorithms that have been used for more than 50 years.

What is your estimate of the impact, which your project may have to society in the long term?

The project will introduce new decision-making tools with numerous potential applications across multiple industries. Some examples include: efficiently orchestrating a team of robots in a warehouse, tailoring assembly lines to create personalized products for each customer, or precisely controlling the functions of a satellite, among others.

The ability of making better decisions in all of these scenarios could have a big impact. Beyond boosting productivity and reducing time and production costs, there is also the potential to reduce energy consumption, helping to make these applications more sustainable.

Which impact do you expect the Sapere Aude programme will have on your career as a researcher?

This Sapere Aude: DFF Starting Grant is very important for my career in two ways. First, it will allow me to develop the idea of Contrastive Analysis to change the way AI systems think about their actions. This is an idea I have been slowly building for years, and now it can be fully explored with the resources this grant provides. 

On the other hand, my medium-term goal is to build a prominent AI planning research group in Denmark. This project allows me to lead a team of two Ph.D. students and a postdoc, paving the way for future growth in the years to come.

Background and personal life

I grew up in Madrid, Spain, where I graduated from the University Carlos III of Madrid. After my PhD, I moved to Saarland University in Germany for 6 years. And finally, I came to Aalborg University in 2020. I really enjoy working in academia, as my main drive is to build knowledge for society and share it with students. 

Apart from research, my hobbies are traveling around the world, chess, and table tennis. But lately, I don't have much time for these activities as my main hobby at the moment is taking care of my child who was born less than a year ago and I'm enjoying the experience very much.