Riccardo Fusaroli
@fusaroli.bsky.social
Cognitive Science at Aarhus Uni. Curious about social interactions, symbolic behaviors, and meta-science. Focus on stats, computational modeling, machine learning, complex systems, language, exp semiotics and neuropsychiatric conditions. He/They.
awesome, thanks!
February 6, 2026 at 2:22 PM
awesome, thanks!
interesting. My last use was trying to feed pathfinder values back into brms, so it took a long time. I'll go back and use cmdstanr directly to try this magic :-)
February 6, 2026 at 2:12 PM
interesting. My last use was trying to feed pathfinder values back into brms, so it took a long time. I'll go back and use cmdstanr directly to try this magic :-)
nice to know about beta-binomials, I'll cover them in my course in 6-7 weeks. As to whether it's worth: it's kind of a hassle to convert the pathfinder outcomes into initial values (or I haven't found a good way yet), so if it saves less than an hour, it's kind of pointless for me :-)
February 6, 2026 at 7:36 AM
nice to know about beta-binomials, I'll cover them in my course in 6-7 weeks. As to whether it's worth: it's kind of a hassle to convert the pathfinder outcomes into initial values (or I haven't found a good way yet), so if it saves less than an hour, it's kind of pointless for me :-)
that's exactly what I'm using :-) I'll share the materials when I get there with the course (in 4-5 weeks I think)
February 4, 2026 at 4:46 PM
that's exactly what I'm using :-) I'll share the materials when I get there with the course (in 4-5 weeks I think)
ahah, I have targets on my list, but I am always so worried that I'll rabbithole there forever :-)
February 4, 2026 at 4:41 PM
ahah, I have targets on my list, but I am always so worried that I'll rabbithole there forever :-)
do you have any tip, selling point? my idea is to make my prior sensitivity checks more systematic and perhaps computationally efficient, but I haven't tried yet
February 4, 2026 at 4:18 PM
do you have any tip, selling point? my idea is to make my prior sensitivity checks more systematic and perhaps computationally efficient, but I haven't tried yet
5. Priorsense (n-kall.github.io/priorsense/) and DeclareDesign (declaredesign.org/r/declaredes...) are on my list to try out :-)
Prior Diagnostics and Sensitivity Analysis
Provides functions for prior and likelihood sensitivity analysis in Bayesian models. Currently it implements methods to determine the sensitivity of the posterior to power-scaling perturbations of the...
n-kall.github.io
February 4, 2026 at 1:37 PM
5. Priorsense (n-kall.github.io/priorsense/) and DeclareDesign (declaredesign.org/r/declaredes...) are on my list to try out :-)
4. Simulation based calibration (mc-stan.org/docs/stan-us...) is a really neat theoretically grounded alternative to parameter recovery.
5/
5/
mc-stan.org
February 4, 2026 at 1:30 PM
4. Simulation based calibration (mc-stan.org/docs/stan-us...) is a really neat theoretically grounded alternative to parameter recovery.
5/
5/
2. quickly iterate for whatever reason: pathfinder (mc-stan.org/docs/referen...) is an amazing quick approximation for HMC.
3. bonus point: for *some* models, using the pathfinder's approximated results as initial values makes the HMC go vrooom! (no idea yet as to a priori which models)
3/
3. bonus point: for *some* models, using the pathfinder's approximated results as initial values makes the HMC go vrooom! (no idea yet as to a priori which models)
3/
Pathfinder
mc-stan.org
February 4, 2026 at 1:30 PM
2. quickly iterate for whatever reason: pathfinder (mc-stan.org/docs/referen...) is an amazing quick approximation for HMC.
3. bonus point: for *some* models, using the pathfinder's approximated results as initial values makes the HMC go vrooom! (no idea yet as to a priori which models)
3/
3. bonus point: for *some* models, using the pathfinder's approximated results as initial values makes the HMC go vrooom! (no idea yet as to a priori which models)
3/
I remember him talking about the appropriation of the semiotic modes of production in this context, which is very fitting :-)
February 4, 2026 at 1:04 PM
I remember him talking about the appropriation of the semiotic modes of production in this context, which is very fitting :-)
darn those pesky null models! ;-)
February 4, 2026 at 12:55 PM
darn those pesky null models! ;-)
very much in line with how Eco described semiotic guerrilla as a reaction to passive media reception: disruptnow.org/disruptive-s...
Towards a Semiological Guerrilla Warfare by Umberto Eco (1986)
Eco argues that controlling the message is less important than controlling the possible interpretations of the message. To disrupt the mainstream mass media, Eco suggests using guerilla tactics, such ...
disruptnow.org
February 4, 2026 at 12:42 PM
very much in line with how Eco described semiotic guerrilla as a reaction to passive media reception: disruptnow.org/disruptive-s...
my youngest has been playing his favorite boardgame with very different comments now...
January 30, 2026 at 8:16 AM
my youngest has been playing his favorite boardgame with very different comments now...
my only speculation was that it was really old japanese, but google lens confirms it's some random chinese :-)
January 23, 2026 at 11:28 AM
my only speculation was that it was really old japanese, but google lens confirms it's some random chinese :-)
the ending was... not what I was expecting... But now we are all eagerly waiting for a second season 🍽️
January 22, 2026 at 2:57 PM
the ending was... not what I was expecting... But now we are all eagerly waiting for a second season 🍽️
the ending was... not what I was expecting... But now we are all eagerly waiting for a second season 🍽️
January 22, 2026 at 2:57 PM
the ending was... not what I was expecting... But now we are all eagerly waiting for a second season 🍽️