Cognitive Limitations of Human Bayesian Learning

\label{fig:fully_rational} Replace this text with your caption

To do list:

  • (Johannes) Explanation rational Bayesian convergence being equal \(\sqrt{t} = t^{1/2}\)
    This follows from the Central Limit Theorem and from the fact that a rationally updated dirichlet conjugate prior acts as a counts model. For details see the following Wikipedia articles: Central Limit Theorem and Dirichlet Distribution.

  • (Thomas) Explain the Hawkes Poisson Process as a candidate model in this context of bounded rationality

  • (Thomas) Check dynamics as a function of 2 minutes sessions \(\rightarrow\) Is there any coherent pattern between spikes and 2 minutes periods

  • (Johannes) Compute matrices of distances between models as a function of time: The output should be triangular matrices

  • (Johannes) Learning models in Economics

\(\frac{Diversity}{\text{Distance from true process}}=C*t^{\alpha - \beta}\), where \(\alpha\) is the rate of decrease in diversity and \(\beta\) is the rate of human or collective learning.