@mic_caprio

Assistant Professor @WarwickDCS. On this profile, math (Imprecise Probabilities/Stats/AI/ML), baseball (@Phillies), and football (@juventusfcen).

 Coventry, England
michelecaprio.wixsite.com/caprio
Joined February 2012
  • Tweets864
  • Following766
  • Followers547
11Media
Excited to help organise #WUML2027 @csmcr on 2 - 4 Feb 2027! Join us for talks, tutorials and posters on uncertainty in ML, from theory to applications. Registration is free, with 90 places available. Speakers and details in my LinkedIn post: lnkd.in/p/eNhwKbXP

WUML 2027: Uncertainty in Machine Learning Workshop | Michele Caprio posted on the topic | LinkedIn

When should a machine learning model express uncertainty, and what should it do when it cannot make a reliable prediction? Since 2017, the Workshop on Uncertainty in Machine Learning (WUML) has...

linkedin.com
2
9
385
Next week I'll be in Edinburgh for workshops.inf.ed.ac.uk/ml/ba… If you're around, let's have a chat!
5
225
What is TMLR doing under a 300% increase in submissions? - Suspensions of >150 authors - Publicly naming major violations - Submission quotas distinguishing slop patterns - Desk rejections 6% → 53%; focus on clarity - AI reviews evaluating soundness blog.tmlr.org/2026/adapting-…

Adapting Our Processes to the Changing World of Scientific Publication

TMLR is adapting its submission and review processes to handle rapid growth while protecting review quality, reviewer time, and scientific integrity.

blog.tmlr.org
4
82
19
640
69,716
Michele Caprio retweeted
sipta@SIPTAtweetsOct 15
The SIPTA Seminars are back! We start the season on October 22 with Davide Petturiti, who will tell us about optimal transport in Dempster-Shafer theory! All the info at sipta.org/events/sipta-semin…
3
3
140
Michele Caprio retweeted
Sam Power@sp_monte_carloOct 14
In between 'leaving' and 'returning to' Twitter over the past however long, I have mostly stopped 'blogging'. Conversely, I have started writing up various bits and pieces as marginally more formal documents, to be found on my website at sites.google.com/view/sp-mon….
2
10
2
66
6,218
Michele Caprio retweeted
Paolo Perrone@PaoloPMathOct 5
It's finally there! arxiv.org/abs/2610.02257

Dagger Categories in Riemannian Geometry

A dagger on a category assigns to every morphism $f \colon X \to Y$ a morphism $f^\dagger \colon Y \to X$, contravariantly and involutively, in the way the transpose reverses a real matrix. We...

arxiv.org
7
20
3
166
15,035
Imprecise Probabilistic Machine Learning (IPML)! For the second time, I’m offering this course in the upcoming winter semester at CISPA Helmholtz Center for Information Security, Saarland University, the University of Hamburg, and online for students around the world. Join us!
2
4
41
1,665
Michele Caprio retweeted
Quanta Magazine
@QuantaMagazineSep 17
In the 1960s, Christopher Zeeman helped popularize catastrophe theory. It was all the rage; people called on it to describe everything from bridge collapses to prison riots to regime changes. The hype soon fizzled. But the underlying concepts didn’t. quantamagazine.org/the-math-…
7
36
191
12,448
Tomorrow I’ll be speaking at #SMPS2026 in Lecce (just 15 miles from my hometown!) about the connections between conformal prediction and imprecise probabilities. Let’s chat if you’re around :) Summary paper: dropbox.com/scl/fi/ti7kg3nmu…

Summary_Paper_Lecce.pdf

Shared with Dropbox

dropbox.com
1
12
339
Michele Caprio retweeted
Stat.ML Papers@StatMLPapersAug 31
Prequential posteriors ift.tt/LH8vF7T

Prequential posteriors

Data assimilation is a fundamental task in updating forecasting models upon observing new data, with applications ranging from weather prediction to online reinforcement learning. Deep generative...

arxiv.org
3
39
3,409
Michele Caprio retweeted
Stat.ML Papers@StatMLPapersAug 27
Gradient-based Sample Selection for Faster Bayesian Optimization ift.tt/CEWvu9k

Gradient-based Sample Selection for Faster Bayesian Optimization

Bayesian optimization (BO) is an effective technique for black-box optimization. However, its applicability is typically limited to moderate-budget problems due to the cubic complexity of fitting...

arxiv.org
2
30
2,055
The LMS are hosting a special session at the 2026 European Women in Mathematics conference at the University of Warwick on Tuesday 1 September from 13:30-18:00, including reception. All conference attendees welcome! For further details ➡️ sites.google.com/europeanwom…
7
21
1,316
"What is mathematics now, and what should it be?" (by Jeremy Avigad): arxiv.org/abs/2608.23218

What is mathematics now, and what should it be?

Advances in neural theorem provers have been impressive, but the successes obscure a broader vision of what AI can do for mathematics and how mathematicians can engage with AI. This essay advances...

arxiv.org
5
16
4
97
10,870
Happy to share a rather unusual new preprint of mine, arxiv.org/abs/2608.15337 My interest was sparked by stimulating discussions with colleagues, that made me think how imprecise probabilities can be beneficial in settling a question in the field of combustion. 1/10
3
6
3
54
3,664
more replies
One aspect of this project I find especially exciting is seeing ideas developed in imprecise probability become useful in resolving an open problem in mathematical analysis, which I guess it's almost a first (the friends from @SIPTAtweets can correct me if I'm wrong). 9/10
1
1
80
Very happy to discuss this first draft! 10/10
1
73
Load more