Thank you! Yeah the results are based on xp (and not protected xp). I’ve also introduced a new method for determining critical values of xp by controlling false positives; the commonly assumed 0.95 critical value for either xp or pxp can be misleading in my view (usually too conservative).

I argue that we need to account for the size of the model space when determining sample size, as larger model spaces reduce power. I also show that the commonly used “fixed effects” model selection approach is statistically unreliable. An analysis of the literature suggests shortcomings in both

More generally, we link MEC coding to planning-ready compositional representations, with invariant and modular responses in ubiquitous MEC object vector cells. These cells provide the building blocks of compositionality in the model.

Neurally, influential work proposed grid cells encode eigenvectors of the successor map. Nice idea, but it struggles when barriers or goals change. Our model ties grid code to the compositional map, keeping them useful even as the world changes, consistent with empirical findings on local remapping.

Computationally, the model builds a successor map piece by piece, by putting together representations related to barriers and goals. We propose translation/rotation-invariant code for representation of task components (objects/goals) that plans near-optimally in complicated navigation tasks.

Great great work, congrats!

So happy for you, congratulations!

So so sorry for you loss. That’s a beautiful piece!

Am I qualified as a cognitive controller?