Adventures in Covariance - Lecture B03 of Statistical Rethinking 2026. Prelude about MCMC effective sample size, Multilevel varying slopes models with correlated features, non-centered parameterizations for covariance models, model expansion workflow. Next week is social networks.
Statistical Rethinking Lecture B03 - Adventures in Covariance
YouTube video by Richard McElreath
www.youtube.com
January 23, 2026 at 1:22 PM
Adventures in Covariance - Lecture B03 of Statistical Rethinking 2026. Prelude about MCMC effective sample size, Multilevel varying slopes models with correlated features, non-centered parameterizations for covariance models, model expansion workflow. Next week is social networks.
the funny thing is that the @lastpositivist.bsky.social tweet is more or less an axiom of AI safety: the larger the covariance matrix, the more ways there are for spurious covariance to get baked into the refusal subspace (such as "poems aren't dangerous"), so best practice is to force a dumb result
November 30, 2025 at 10:52 PM
the funny thing is that the @lastpositivist.bsky.social tweet is more or less an axiom of AI safety: the larger the covariance matrix, the more ways there are for spurious covariance to get baked into the refusal subspace (such as "poems aren't dangerous"), so best practice is to force a dumb result
Excited to share a paper recently accepted at Nature Mental Health! 😄
For decades, we've assumed that covariance among psychopathology informs comorbidity. What if we’ve been wrong?
For decades, we've assumed that covariance among psychopathology informs comorbidity. What if we’ve been wrong?
Covariation among higher-order psychopathology dimensions is driven by nomorbidity, not comorbidity: https://osf.io/3t2cm
July 9, 2026 at 12:21 AM
Excited to share a paper recently accepted at Nature Mental Health! 😄
For decades, we've assumed that covariance among psychopathology informs comorbidity. What if we’ve been wrong?
For decades, we've assumed that covariance among psychopathology informs comorbidity. What if we’ve been wrong?
Gaussian approximation of a target distribution: mean-field versus full-covariance! Below shows a simple gradient descent on KL(q||p)
November 20, 2024 at 8:51 AM
Gaussian approximation of a target distribution: mean-field versus full-covariance! Below shows a simple gradient descent on KL(q||p)
This was always where transphobia was gonna end up, the covariance on these two bigotries is crazy high
March 28, 2026 at 11:07 AM
This was always where transphobia was gonna end up, the covariance on these two bigotries is crazy high
Surgeonfishes are cool! Our Evolution paper now out in final form, the amazing patterns of change in fish shape and covariance (new method!) elegantly revealed by @linnealungstrom.bsky.social
🐟🐠🦑🌿
academic.oup.com/evolut/artic...
🐟🐠🦑🌿
academic.oup.com/evolut/artic...
Phylogenetics, trait covariance analysis, and the evolution of fin and body shape in the surgeonfishes
Abstract. Insights into ecomorphology are enriched when investigating evolutionary correlations between ecologically and functionally important traits. How
academic.oup.com
August 16, 2026 at 12:40 AM
Surgeonfishes are cool! Our Evolution paper now out in final form, the amazing patterns of change in fish shape and covariance (new method!) elegantly revealed by @linnealungstrom.bsky.social
🐟🐠🦑🌿
academic.oup.com/evolut/artic...
🐟🐠🦑🌿
academic.oup.com/evolut/artic...
so who's excited about the lme4 getting covariance matrices
January 7, 2026 at 7:51 PM
so who's excited about the lme4 getting covariance matrices
Cory Booker now giving one of the best explanations of covariance/contravariance I've ever seen, how is he still lucid?
April 1, 2025 at 10:43 PM
Cory Booker now giving one of the best explanations of covariance/contravariance I've ever seen, how is he still lucid?
I've always appreciated the beauty of hot air balloons, but after learning about atmospheric physics for eddy covariance theory I'm amazed by the science at work.
🧪
🧪
March 18, 2025 at 8:56 PM
I've always appreciated the beauty of hot air balloons, but after learning about atmospheric physics for eddy covariance theory I'm amazed by the science at work.
🧪
🧪
Week 8 of my Bayes stats course, now deep in applications of multilevel covariance modelling: social networks and Gaussian processes. Everything wiggles in large amounts. Lecture: www.youtube.com/watch?v=Y2ZL...
February 28, 2025 at 9:26 AM
Week 8 of my Bayes stats course, now deep in applications of multilevel covariance modelling: social networks and Gaussian processes. Everything wiggles in large amounts. Lecture: www.youtube.com/watch?v=Y2ZL...
Walking towards the eddy covariance tower at Poker Flats, 40 minutes’ drive northeast of Fairbanks, AK
July 29, 2026 at 5:13 AM
Walking towards the eddy covariance tower at Poker Flats, 40 minutes’ drive northeast of Fairbanks, AK
A Painstakingly long post on covariance and how projection matrices are ESSENTIAL for understanding these things.
And look! Interactivity!
And look! Interactivity!
May 14, 2026 at 10:48 PM
A Painstakingly long post on covariance and how projection matrices are ESSENTIAL for understanding these things.
And look! Interactivity!
And look! Interactivity!
like the basic idea is that everything necessary to make a god is already present in the covariance structure of all human language. which, lol
November 8, 2025 at 6:54 PM
like the basic idea is that everything necessary to make a god is already present in the covariance structure of all human language. which, lol
Covariance is coming to Literal.
November 13, 2024 at 11:18 AM
Covariance is coming to Literal.
I will never forget the groans Niels Waller elicited by teaching “A correlation is just a covariance computed in France—between z scores!”
August 25, 2025 at 4:09 PM
I will never forget the groans Niels Waller elicited by teaching “A correlation is just a covariance computed in France—between z scores!”
I take statistics very seriously
January 23, 2026 at 8:28 AM
I take statistics very seriously
November 23, 2025 at 2:30 PM
I am always surprised that scientists are unaware of Ledoit and Wolf, which to my mind is the coolest generally applicable result of the quant finance research program: shockingly, it is possible to optimally estimate a covariance matrix from very underdetermined data events.bse.eu/live/files/2...
events.bse.eu
March 20, 2026 at 12:33 AM
I am always surprised that scientists are unaware of Ledoit and Wolf, which to my mind is the coolest generally applicable result of the quant finance research program: shockingly, it is possible to optimally estimate a covariance matrix from very underdetermined data events.bse.eu/live/files/2...
I had a great time writing this dispatch @currentbiology.bsky.social about Jacob Suissa's recent paper on the link between fern phyllotaxis and vascular architecture. "Evo–devo: Ferns flourished due to developmental covariance between leaves and vasculature"
authors.elsevier.com/a/1mH793QW8S...
authors.elsevier.com/a/1mH793QW8S...
December 15, 2025 at 4:09 PM
I had a great time writing this dispatch @currentbiology.bsky.social about Jacob Suissa's recent paper on the link between fern phyllotaxis and vascular architecture. "Evo–devo: Ferns flourished due to developmental covariance between leaves and vasculature"
authors.elsevier.com/a/1mH793QW8S...
authors.elsevier.com/a/1mH793QW8S...
Brain Microstructural Alterations in Children Post‐COVID‐19 Infection Through VBM, SBM, and Structural Covariance Network Analysis - Wang - 2025 - Brain and Behavior - Wiley Online Library onlinelibrary.wiley.com/doi/10.1002/...
Brain Microstructural Alterations in Children Post‐COVID‐19 Infection Through VBM, SBM, and Structural Covariance Network Analysis
Background Children represent a particularly vulnerable group to the long-term consequences of COVID-19 due to their ongoing neurodevelopment. This study aimed to identify transient and persistent s...
onlinelibrary.wiley.com
September 23, 2025 at 9:12 PM
Brain Microstructural Alterations in Children Post‐COVID‐19 Infection Through VBM, SBM, and Structural Covariance Network Analysis - Wang - 2025 - Brain and Behavior - Wiley Online Library onlinelibrary.wiley.com/doi/10.1002/...
Using gene–gene covariance structure to represent cellular neighborhoods improves the integration of single-cell RNA sequencing and spatial transcriptomics data go.nature.com/4auQI72
rdcu.be/ebBiU
rdcu.be/ebBiU
The covariance environment defines cellular niches for spatial inference - Nature Biotechnology
Using gene–gene covariance structure to represent cellular neighborhoods improves the integration of single-cell RNA sequencing and spatial transcriptomics data.
go.nature.com
March 13, 2025 at 10:14 PM
Using gene–gene covariance structure to represent cellular neighborhoods improves the integration of single-cell RNA sequencing and spatial transcriptomics data go.nature.com/4auQI72
rdcu.be/ebBiU
rdcu.be/ebBiU
Statistical Rethinking 2026 Lecture B01 Multilevel Models is online. This is the first lecture of the "experienced" section, in which we start with multilevel models and venture into vast covariance spaces. Full lecture list still here: github.com/rmcelreath/s...
Statistical Rethinking 2026 - Lecture B01 - Multilevel Models
YouTube video by Richard McElreath
www.youtube.com
January 9, 2026 at 10:34 AM
Statistical Rethinking 2026 Lecture B01 Multilevel Models is online. This is the first lecture of the "experienced" section, in which we start with multilevel models and venture into vast covariance spaces. Full lecture list still here: github.com/rmcelreath/s...