Computational Neuroscience and Education Science

Artificial Intelligence Devoted to Education (AIDE):

We want to explore to what extent approaches or methods from cognitive neuroscience, linked to machine learning and knowledge representation, could help to better formalize human learning as studied in educational sciences. In other words: we are taking advantage of our better understanding of how our brains work to help us better understand how our children learn. The focus here is on learning computational thinking, i.e., what to share in terms of skills needed to master the digital world, not just consume or endure it. To achieve this, we consider modeling specific learning tasks involving creative problem-solving and tangible robotic artifacts.

General presentation

Preliminary outcomes

Contributors

Contact

French summary

  General presentation

 

 

The three operational objectives of the AIDE project

• At the theoretical level, using machine learning formalism to model the learner, as it can be done in computational neuroscience on another scale, e.g. considering reinforcement learning framework to account for human learning behaviors.
• At the experimental level, using machine learning tools to obtain more formalized, more reliable, and more automated measures of variables related to a learning situation, using real-life objects (e.g., unplugged activities, manipulation of tangible connected objects).
• At the pedagogical level, helping the learner understand their own learning process, making explicit the mechanisms involved in creativity and problem-solving.

In other words, before using artificial intelligence to improve human learning, let’s try to use it to better understand how humans learn.

This Exploratory action (AEx) is a four-year project that started end of 2020 and until 2024.

  Preliminary outcomes

We introduced the idea of a symbolic description of a complex human learning task, in order to contribute to better understand how we learn [mercier:hal-03324136], in the specific context of a task, named  #CreaCube, related to the initiation to computational thinking presented as an ill-defined problem. We assume that such a task engages the subject in solving a problem while appealing to creativity.

We proposed to apply a reinforcement learning paradigm to a symbolic representation of a concrete problem-solving task, formalized by an ontology [mercier:hal-03327706, mercier:hal-03791482]. This work is embedded in strong collaboration with education science collaborators [romero:hal-02957270] working on computational thinking initiation and computer science tools in education with a multi-disciplinary vision of cognitive function modeling [denet:hal-03382314].

We also proposed to map an ontology onto a SPA-based architecture with a preliminary partial implementation into spiking neural networks [mercier:hal-03360307, mercier:hal-03550354], in order to provide an effective link between symbolic knowledge representation and biologically plausible numerical implementation, also considering uncertainty by revisiting the possibility theory [vallaeys:hal-03338721].

Contributors

• Axel Palaude (PhD Student)
• Chloé Mercier (PhD Student)
• Frédéric Alexandre (Researcher)
  
• Margarida Romero (Researcher)
• Thierry Viéville (Researcher)

with the advice and help of

Gérard Giraudon (Researcher), Dorian Mazauric (Researcher, adviser), Éric Pascual (Engineer, adviser), Didier Roy (Researcher, adviser), Hélène Sauzéon (Researcher, adviser), and thanks to former collaborators Sabrina Barnabé (Craftswoman), Jia Fu (former Intern student), Lola Denet (former MSc Intern student), Gabriel Doriath-Döhler (former BSc Intern student), Thalita Firmo-Drumond (former PhD student), Ali Issaoui (former Intern Student), Younes Jallouf (former MSc Intern student), Divya Menon (former MSc Intern student), Bhargav Teja Nallapu (former PhD student), [É]Lisa Roux (former PostDoc student), Théophane Vallaeys (former BSc Intern student).