Project
Cascading Transitions in Learning

The goal is to build formal models of learning both at the individual level
and social network levels using tools from complexity science and
agent-based modelling. At the individual level we developed a reciprocal
model of practice and success in free-learning (such as hobbies like Chess)
[1], integrating previous findings on learning curves and forgetting into a
formal model. Empirically, we will use real-world data from adaptive
learning platforms such as Lichess or Prowise Learn [2].

In the future we will review known phenomena on peer effects and further
model how learning, interest and curiosity spread socially. An example of
the approach is the HIOM model in opinion dynamics [3] which extended a
previously developed network model of attitudes within individuals [4].

Why IOPS Project: The project involves formal modelling using language of
differential equations and stochastic processes. Furthermore, network
based ABMs will be used to simulate learning, motivation or curiosity at the
social level. We also use empirical datasets from adaptive learning and
playing platforms such as Lichess and Prowise-Learn where ELO-type
algorithms (based on IRT models) are used to track changing abilities and
learning over time. All these make it very relevant for IOPS.

Supervisors
Prof. Dr. Han van der Maas
Dr. Alexander Savi
Dr. Brenda Jansen

Financed by
ERC Grant: “Beyond the Tipping Point: Cascading Transitions in the
Behavioral and Social Sciences” (ERC, Cascade, 101053880.)

Period
1 April 2023 – 1 April 2027