RELEARN · Goal-directed learning of the statistical structure of the environment
Horizon 2020 — Marie Skłodowska-Curie Actions
- Duration
- 2021-01-01 → 2022-12-31
- EU contribution
- €174,806
- Participants
- 1
- Scheme
- MSCA-IF
Lines connect the coordinator with its partners.
Results in brief
Goal-directed learning of the statistical structure of the environment
How humans learn representations of their environment is a question to which cognitive science and neuroscience offer only very fragmentary answers currently. Existing theoretical frameworks don’t incorporate the element of information-directed exploration into the representation learning process, and experimental paradigms don’t aim to directly assess representational dynamics through a behavioural lens. Advancements on both of these fronts are required to produce a useful model of human task-based representation learning, that is able to make predictions about behaviour observable on the short time scales of practical experiments in such a way that allows for the empirical evaluation of algorithmic hypotheses. We need to reconcile theoretical approaches addressing separate aspects of this problem, such as Bayesian inference and model-based reinforcement learning, as well as extend them with an explicit consideration of meta-cognitive decision making about perceptual compression. Existing experimental paradigms have to be extended such that they present naturalistic decision-making problems to participants and allow for the evaluation of learned representations both behaviourally and neurally using imaging methods. The action produced both experimental advances in the form of two novel experimental paradigms, and theoretical ones, including a more complete understanding of the relationship between the existing frameworks of Control as Inference and the Information Bottleneck, and most importantly, it opened the entirely new research direction of representational planning, with implications in both machine learning and cognitive science.
Data: CORDIS, © European Union
Project objective
Learning the statistical buildup of the environment serves the purpose of making good decisions, thus what regularities humans learn and what ones they neglect depends on the relevance towards maximizing reward. Recent studies characterise reward-based modulation of feature representations built by humans and animals both on the behavioural and neural level, but the effect of reward on learning higher-order environmental statistics is unknown. Our hypothesis is that humans do not learn to represent feature co-occurrence statistics if it does not help to predict reward due to resource constraints on computation and storage. We propose a mathematical framework based on Bayesian hierarchical modelling and reinforcement learning to predict the modulatory effect of reward on learned representations. We will test the predictions of the model in a series of experiments where humans need to learn to associate precisely controlled statistical aspects of a naturalistic simulated environment to reward both in the lab and online, in reactive and planning-based tasks. Additional to behaviour, the model will predict the structure of neural representations and their changes over the course of the experiment as well. We will test those predictions using magnetoencephalography during the learning phase of the experiments and decoding analysis to compare model variables to neural responses. The results will contribute to the understanding of representational learning in humans, with potential implications in psychiatry and economics as well as supply the community with novel analytical tools and data. The unique mentoring at the host institution together with the extensive training program including international visits to world-leading collaborators will establish my independent research program in computational neuroscience.
Original text from CORDIS.
Participants
- MAX-PLANCK-GESELLSCHAFT ZUR FORDERUNG DER WISSENSCHAFTEN EV · MUNCHENCoordinatorGermany
Links
Data: CORDIS, © European Union