Riccardo Fusaroli

yeah, thanks! still working on proper suppressions and keeping graphics consistent across chapter, etc.. i promised my students I would upload them before the lecture, so I was mostly concerned in making everything work :-) At the end of the course I have plan to turn it into quarto as well.

Actually what is needed is to change the experimental setup, as this is a memory model playing against a biased agent. Both good memory and a pre-existing bias could underlie the agent's (acquired) bias. We need to change the opponent (e.g. to reversal learning) Nice pedagogical point!

it will disappear with more data, I thought (or well didn't think hard enough, so not sure it could be called "thought") 3/

the culprit is ofc correlation btw parameters. 2/

as usual just the act of posting made me find a bunch of issues :-) adding now a favorite bit: in one of the models the parameters are not nicely recovered: 1/

my course notes on a bayesian workflow for (single agent) cognitive modeling are now fully revised and online: fusaroli.github.io/AdvancedCogn...

Predictive checks, updating checks, sensitivity analyses and simulation based calibration in @mc-stan.org

Feedback is very welcome!

The Silicon Valley Wager: If the AI God will be real, the reward will be infinite. If it's not real, you will meet tragic, but ultimately finite losses. Therefore, the only rational choice...

🚀 Postdoc Alert! Are you passionate about social learning & cultural evolution? @dominikdeffner.bsky.social & I have a 3-year position with freedom to develop your research and work on cutting-edge multiplayer and immersive experiments. Apply by March 30! hmc-lab.com/SocialLearni... Pls share 🙏

Female friendship and the cultural transmission of low-fertility values: Evidence from rural Poland Abstract. Understanding how demographic and cultural change interact is essential for understanding the trajectory of historical and contemporary societies

This: academic.oup.com/pnasnexus/ar...

I noticed it while watching new editions of @rmcelreath.bsky.social 's stat rethinking course as well. They are terrific for me, I always learn something new and refine my conceptions. Tho' I lack the empirical data as to whether new watchers find them as intuitive as the earlier versions.

ahah, now that you mention attention/memory I feel like I want a set of different cognitive profiles in the algorithm (the narrow-focused, the explorer, etc.). cool work!

Stan for cognitive modeling - cheatsheet Role & Persona You are a distinguished specialist in Bayesian cognitive modeling and computational statistics, applying probabilistic models to experimental and observational data. You operate with un...

hmm, here are my best practices organized in skill-like format (well, I use gemini gems, so it might be a bit different): docs.google.com/document/d/1...

eh, good for a first draft, but I'm still a fan of manual curation :-)

Teaching my cognitive modeling every year, I keep updating my contents. Clearer explanations (I think), fewer factual errors, more precisions. Also, new applications burge in, more updated methods, and more explicit connections across parts. Unsure as to whether that overloads some students...

the rants are what keeps me coming back to the new editions of the courses 🤷

This is an interesting discussion. While a Bayesian at heart, I find that most technical things can be done with either approach if you're skilled enough, but the different approaches make different things easier. Bayes4TW w convergence, priors, workflows and top reason bespoke modeling.

Does anyone know an example (paper, blog post) where frequentist and Bayesian methods lead to different conclusions?

#rstats #stats

Or an anecdote that lead you to embrace one or the other approach?

1) building a multilevel model and extracting individual estimates to be then correlated w relevant symptoms; vs 2) building the correlation already in the multilevel model. 1 leads to correlations all over the place, 2 to fewer and more sensible ones.

Taking others into account: combining directly experienced and indirect information in schizophrenia Simonsen et al. report that hallucinations, delusions and negative symptoms in schizophrenia may stem from a common underlying abnormality in inference, wh

Hmm, to me the sovereign advantage is the ability to model your assumptions/theories in ways that is not obvious with frequentist techniques (I guess I'd have to build my own estimator). In this paper - academic.oup.com/brain/articl... - we compared something doable in freq, wi something less so 1/

Against theory-motivated experimentation: Can random experimental choice lead to better theories? - Marina Dubova, Arseny Moskvichev, Kevin Zollman, 2026 Scientists must choose which among many experiments to perform. We study the epistemic success of experimental choice strategies proposed by philosophers of sci...

cool paper on the trade-offs between theory-driven and random research: journals.sagepub.com/doi/10.1177/... (by @mdubova.bsky.social @kevinzollman.com). It made me think of @devezer.bsky.social's wise words on assessing the assumptions of research policies. More thoughts when I read it fully :-)

this looks pretty useful even just as a cheatsheet to give students. Anybody has done something similar for @mc-stan.org ? like a highly distilled stan user's guide.

📣📣📣

Applications for the 2026 Diverse Intelligences Summer Institute (DISI) are now open!!

Are you interested in intelligence, mind, and cognition in all its forms? Early-career scholars from any discipline—and storytellers in any medium—are encouraged to apply!

More info: disi.org

*Preprint*: Pedagogy in the speech-gesture couplings of caregivers: Evidence from a corpus-based analysis by @marinewang.bsky.social, @eddonnellan.bsky.social and myself: doi.org/10.31234/osf...

Progress bar showing 2800/8000 35% ETA: 4m

Help to test new progress bar in CmdStanR by @josswright.bsky.social, see more at discourse.mc-stan.org/t/help-us-te...

✨New Perspective out w/ Wenning Deng and @fearbrain.bsky.social in @cp-iscience.bsky.social ! We argue that social foraging gives us a unifying, and ecologically grounded way to study how decisions unfold across levels — from individuals and dyads to collectives.

www.cell.com/iscience/ful...

Cognitive primitives of the insect brain Understanding the mechanistic basis of human cognition is likely to benefit from investigating how it emerged through evolution. We propose that identifying and investigating fundamental brain functio...

Little smart critters..🪰🐝🐜
Cognitive primitives of the insect brain: Trends in Cognitive Sciences www.cell.com/trends/cogni...

🚨 New Paper: How can AI help us understand child lang dev? If we train models on children’s environment, they can tell us if this environment support learning.
E.g., models tested child linguistic input (Huebner et al.) and visual input (Vong et al.).

What about Social Interaction? (a thread 🧵)

Humanity's oldest geometries, engraved on ostrich eggs At several archaeological sites in southern Africa, hundreds of highly unusual fragments of ostrich eggs have been found. Dating back more than 60,000 years, the shells were engraved by groups of Homo...

While everyone is excited by early European precursors of 'writing', let's not forget the African geometrics, here Howiesons Poort oes, but also Blombos haematite, almost twice as old

phys.org/news/2026-02...

and here a few attempts at better understanding how these patterns might or not have been shaped by different cognitive/cultural constraints:
- www.pnas.org/doi/10.1073/...
- onlinelibrary.wiley.com/doi/10.1111/...
- www.sciencedirect.com/science/arti...

this signal detection based missing data estimation model (early stage) is brought to you by mu-ko the happy cow I'll have to work around for the next six month or so (or until my kids come around again and draw the next nuis... err masterpiece)