Thomas Hamelryck

Professor of machine learning. Structural bioinformatics, deep probabilistic programming. Grumpy Girardian. Opinions are my own.

[Thomas Hamelryck @THamelryck]

@RasmusJarlov @USAmbDenmark It's an even lonelier job thinking the US is an outright "enemy". The mind boggles.

[Thomas Hamelryck @THamelryck]

@grantmaxwell Like Nietzsche, Deleuze is a prophet of romanticized delirium, yearning for the transgressive sacred as an antidote to stagnant neurosis. Well worth a read.

[NeurIPS Europe Conference @EurIPSConf]

FAQ:

NeurIPS Europe is an official NeurIPS 2026 satellite event taking place in Paris, France, alongside the main conf in Sydney and the other satellite event in Atlanta.

NeurIPS authors can present their papers at any of the three locations, subject to space availability.

[Andrew Gordon Wilson @andrewgwils]

This is your annual reminder that we don’t need to speculate about whether we will have a “theory of deep learning” and what form it might take, because we already have a basic understanding of generalization in deep learning: https://t.co/AgHdSQjCvU

[Thomas Hamelryck @THamelryck]

@WTSmith17 Lacan, Derrida, Zizek...not quite nonsense, but not conventional rationality either. They're akin to Zen masters, tricksters... They wield a carnivalesque Skillful Means / Upaya, so to say. It didn't go down well under your critical questions :)

[marek mutwil 🇺🇦 @LabMutwil]

🌿 https://t.co/xr5bszJp1j just got a refresh — organised around our three research pillars (gene function prediction · stress resilience · evolution of regulatory circuits), with all our open tools, papers, and team in one place.

And we're hiring 5 PhDs + 1 postdoc 👇

[Thomas Hamelryck @THamelryck]

"AlphaFold's Bayesian roots in Probability Kinematics," co-authored with directional statistics pioneer Prof. Kanti V. Mardia on his 90th birthday, is now published by AISTATS 2026.

In 2018, CASP13 (the 13th edition of the "olympic games" of protein structure prediction) reported "[...] unprecedented progress in the ability of computational methods to predict protein 3D structure. The reasons are not yet fully clear [...]". DeepMind's AlphaFold, version 1, had changed protein structure prediction forever!

AlphaFold1 was published as “Improved protein structure prediction using potentials from deep learning” by Nature in 2020. It uses deep networks to parameterize potentials over dihedral angles and pairwise distances, combining the two using a third reference potential. This approach was heuristically justified by referring to the classic work by the physicist John Kirkwood from 1935 on potentials of mean force for liquids. But there’s more to the story, and it builds on some fascinating work done well after 1935.

First, AlphaFold1 formulates a potential that corresponds to a prior distribution over protein dihedral angles (phi, psi) using directional statistics (which deals with angles, directions, orientations and so on). This goes back to the classic work of David L. Dowe at the end of the 90ies, and our joint work with Kanti V. Mardia around 2010 on generative models based on directional Bayesian networks.

Second, a non-local potential concerning pairwise distances D between amino acids is added. Why? In principle, the dihedral angles are enough to parameterize the 3D structure, but small errors quickly propagate, rendering them only accurate on a local scale.

Third, a reference potential on D is subtracted. Why? Correctly combining distance and angular potentials requires taking into account the signal concerning distances already present in the angular potential. This approach was pioneered as a heuristic by Manfred Sippl in the 1990s, and in the 2010s formalized by us as a surprising application of generalized Bayesian updating called probability kinematics or Jeffrey’s updating (referring to the probability theorist Richard C. Jeffrey, 1926-2002).

AlphaFold1 powered the whole probabilistic setup based on directional statistics and probability kinematics outlined above with deep networks, obtained a maximum a posteriori estimate using simple gradient descent, and subsequently aced CASP13.

Our article connects a paradigm shift in protein structure prediction with the classic work by Richard C. Jeffrey (1950s, probability kinematics), Manfred Sippl (1990s, knowledge based potentials), and our joint work with Kanti V. Mardi (2010s, probability kinematics applied to directional priors), revealing generalised Bayesian updating as a potentially powerful method to formulate compositional deep models from simpler components. The figure shows the AlphaFold1 potential reformulated as a joint probabilistic model.

Article:
https://t.co/Vz3w5ABSyZ

[Thomas Hamelryck @THamelryck]

@PeterSjostedtH Or a God beyond myth, rather.

[Thomas Hamelryck @THamelryck]

@PeterSjostedtH Heraclitus: ‘Dionysos is the same thing as Hades.’

[Thomas Hamelryck @THamelryck]

@antoniogm @FukuyamaFrancis The title is NOT "The end of history" to boot 🤭

[Thomas Hamelryck @THamelryck]

@Grady_Booch I hear AN Whitehead 😎

[Thomas Hamelryck @THamelryck]

Give this man a cigar. LLMs are heuristic induction only - no deduction, no abduction (ie. model building based on a world representation, including causality and memory). They are eminently useful but not conscious, sentient or intelligent, and they will never be.

[Big Brain AI @realBigBrainAI]

Oxford AI professor Michael Wooldridge: "ChatGPT doesn't understand anything. It's essentially doing some fancy statistics."

[Thomas Hamelryck @THamelryck]

@docmilanfar Blatant mediocrity is what it takes to succeed, rather.

[Thomas Hamelryck @THamelryck]

@GaryMarcus Google 1 - Anthropic 0

[Thomas Hamelryck @THamelryck]

@BiologyAIDaily Fragments! Back to the future?

[Thomas Hamelryck @THamelryck]

@magattew Indeed! Imagine the enormous potential. 🏆

[Thomas Hamelryck @THamelryck]

@NeuroTechnoWtch Sure, consciousness can be reduced to running some linear algebra at scale on some gaming chips 🫢

[marek mutwil 🇺🇦 @LabMutwil]

🇩🇰 Built a tool to help find research collaborators across Denmark.

1,300+ faculty from KU, AU, DTU, CBS, SDU. Describe your project in plain English → Claude ranks the best fits with a "why" for each.

Free, public. https://t.co/OPuP3G9IR6

[Thomas Hamelryck @THamelryck]

The world of Bayesian statistics looks upon the terminology drift in the world of machine learning with increasing horror...

[Andrew Gelman et al. @StatModeling]

In ML, everyone’s Humpty Dumpty
https://t.co/q6RB5DFd9W

[Thomas Hamelryck @THamelryck]

@positivememes @Girardism The Buddhists know this very well.