Here’s a link to the paper again: There’s a lot more in there, including an investigation of different learning regimes. Thanks for reading and please feel free to get in touch with any questions or if you want to discuss! (11/11)

Then, we trained artificial neural networks to perform this kind of task and asked what kind of representation they develop! (4/11)

We focused on modularity in contextual tasks, since they are widely used in neuroscience and typically have a nice structure: something like linear classification within a context, but the whole task is complex and not linearly separable in the experimental variables. (3/11)

Inspired by neuroscience results showing both the presence and absence of single-neuron and population-level structure in neural responses, we wanted to understand what conditions give rise to these different representations. (2/11)

RT @BrettMmurphy: Leaked docs: While Gaza starved, USAID and State's refugees bureau came to a legally explosive conclusion — Israel had de…

RT @djfreedman: New FreedmanLab paper shows that the primate superior colliculus is strongly and causally engaged in abstract cognition. ht…

RT @NatureNeuro: The macaque brain can link action to value in a choice task through neural population subspaces, balancing reliable bindin…

Thanks for reading and feel free to get in touch if you have any questions!

I'll also take the opportunity to note that @allemanjm has been an amazing collaborator and is starting to look for postdocs -- I'm sure he'll be an amazing postdoc too. Get in touch with him if you want him to be *your* amazing postdoc!