Rewards and Emotions as Incentives for Collaboration

CY Initiative: Could you introduce yourself and tell us more about your career and your main research themes? 

I'm a computer scientist by training, and I teach in the computer science department at CY Cergy Paris University. I hold a position that places more emphasis on research than teaching, known as a junior professorship. 

I did my thesis at CY Cergy Paris University in the ETIS laboratory, where I'm currently a researcher. It focused on neuroinspired robotics, a discipline that involves designing algorithms inspired by brain function and implementing them in robots. In my thesis, I demonstrated the value of considering emotion and integrating affective processes into these algorithms to make a robot's behavior more intelligent and adaptive.   

I then had two postdoctoral experiences where I focused my research more on neuroscience and in particular on problems linked to decision-making.   

My first post-doc was at Sorbonne University, where I used computational models to analyze experimental data in mice. This enabled me to highlight the ability of mice to increase the variability of its decisions, and to make quasi-random sequences of choices to obtain greater rewards.  

During my second post-doctorate, at the Italian Institute of Technology in Genoa, I was able to combine my modeling work, which is more or less my core business, with setting up experiments to study human behavior in human-robot interactions and brain activity when interacting with a humanoid robot. This research has shown that the robot's non-verbal behavior, such as gaze or gestures, can be perceived by humans as social behavior and can also influence human learning and decision-making processes.  

Today, I link these two subjects as my research in the ETIS laboratory focuses on the interface between neuroscience and robotics. My aim is to gain a better understanding of mental processes linked to affect and decision-making in individual contexts or social interactions.   

CY Initiative: You won the CY Initiative's 2024 call for proposals. What does this project involve, and what is its objective? 

The aim of the JAVA project is to study decision-making in the context of joint action. Joint action occurs when two individuals need to coordinate their actions to achieve a common goal. When we study decision-making in joint action, we aim to determine the extent to which collaborative behavior is governed by decision-making mechanisms that can be said to be fundamental, linked to reward processing or emotion.    

In this project, the idea is to take an interdisciplinary approach and combine the different methodologies I've come to know through my career. This means combining human-human joint action experiments with computational model analysis, EEG (electroencephalography) recording of brain activity, and human-robot experiments.  

There is a growing interest in issues related to joint action in various research communities, notably in cognitive science, psychology and human-robot interaction. From the point of view of scientific literature, this is a very topical issue. Acting jointly with another individual is a ubiquitous action in our social life, in our social experience as human beings. We can find an infinite number of examples in a multitude of contexts to illustrate this fact. Take, for example, a group of firefighters: their coordination is vital to putting out a fire.  

If we want to work towards a common goal, we also need to make decisions that take our partners into account. In this decision-making process, we know that decision-making in both humans and animals generally relies on affective processes, including reward and emotional processing. But in the end, these aspects are relatively little studied in existing work on joint action. 

CY Initiative: In concrete terms, how will the project be rolled out? 

Above all, the project aims to respond to a fundamental research challenge linked to understanding the cognitive and cerebral mechanisms at the heart of an important aspect of human social cognition: the ability to collaborate and coordinate decisions.  

With this kind of approach, the idea is to be able to characterize and formalize things, and achieve what we might call a “mechanistic understanding”. It's about understanding mechanisms and arriving at theoretical models defining how people make decisions in certain contexts, and what factors may influence these decision-making mechanisms.   

In this project, we will study these mechanisms in a specific context to create models that could be crucial for clinical applications and understanding mental disorders. From the moment we have a formal, theoretical characterization of mechanisms that are applied in a mental process, such as decision-making, we can then extend this study to subjects from certain clinical populations, such as schizophrenia, depression, etc. Being able to characterize things in this way can ultimately enable us to identify a little more precisely the deficits or mechanisms that may be dysfunctional in certain disorders, and thus enable us to deploy research that is a little more applied in the clinical sense. At this stage, however, this is not the focus of the project, but a possible use for the future.   

In concrete terms, we have launched an initial experiment in which we collected data with humans. We will soon be starting a new phase of experimentation, this time with robots. These experiments are purely behavioral. We've given participants a task to do with another person or a robot, in order to analyze their behavior, particularly the decisions they make.  

The next step will be to analyze the participants' brain activity during our experiments. The Emergence program provides us with the means to acquire a recording EEG system. 

CY Initiative: The project also focuses on a human-robot angle. How does this come about? 

In order to better understand collaborative behavior, we're going to look at how humans interact with robots, i.e. machines, artificial devices. Robots are destined to play an increasingly important role in our everyday environment. The question is therefore an important one today. 

The aim is to observe the potential differences in people's interactions depending on whether they are facing another human being or a robot. On the one hand, this will enrich our theoretical understanding of social cognition in humans, but it will also enable us to observe certain factors that may influence the way in which individuals interact more or less well with a robot.   

It's important to understand how we humans will respond to these increasingly complex and present machines. In fact, the CNRS ethics committee recently published a report on the importance of studying these so-called “social” robots. A project like JAVA will not only provide some answers through the results of these human-human comparisons, but will also help develop a methodology for studying and analyzing situations where humans work with robots.

Learn more

• Presentation of ETIS lab