iiinova — where your digital self keeps evolving

Daniel Csillag @dccsillag.xyz · 11/07/2025

Off to ICML soon! Excited to talk all things ML&stats: UQ, e-values, safety, optimization, etc If you'd like to meet, shoot me an email!

000

Daniel Csillag @dccsillag.xyz · 24/04/2025

I'll be presenting Strategic Conformal Prediction at AISTATS next week. Looking forward to chatting about all things uncertainty quantification!

010

Daniel Csillag @dccsillag.xyz · 04/02/2025

Happy to announce that our paper, 𝗦𝘁𝗿𝗮𝘁𝗲𝗴𝗶𝗰 𝗖𝗼𝗻𝗳𝗼𝗿𝗺𝗮𝗹 𝗣𝗿𝗲𝗱𝗶𝗰𝘁𝗶𝗼𝗻, is now accepted to AISTATS 2025! Is your uncertainty quantification robust to people trying to break it? Ours is :)

Front page of our paper, 'Strategic Conformal Prediction'.

Abstract: When a machine learning model is deployed, its predictions can alter its environment, as better informed agents strategize to suit their own interests. With such alterations in mind, existing approaches to uncertainty quantification break. In this work we propose a new framework, Strategic Conformal Prediction, which is capable of robust uncertainty quantification in such a setting. Strategic Conformal Prediction is backed by a series of theoretical guarantees spanning marginal coverage, training-conditional coverage, tightness and robustness to misspecification that hold in a distribution-free manner. Experimental analysis further validates our method, showing its remarkable effectiveness in face of arbitrary strategic alterations, whereas other methods break.

110

Reposted by Daniel Csillag

Andrew Gordon Wilson @andrewgwils.bsky.social · 03/01/2025

We're excited to announce the ICML 2025 call for workshops! The CFP and submission advice can be found at: icml.cc/Conferences/.... The deadline is Feb 10. Submit some creative proposals!

icml.cc

ICML 2025 Call for Workshops

01510

Reposted by Daniel Csillag

uai2026 @auai.org · 03/12/2024

The 41st Conference on #Uncertainty in #AI will be held in Rio de Janeiro 🇧🇷, July 21-25! The CfP is out 👉 www.auai.org/uai2025/call... 🚨 Feb 10: Paper submission 🗣️ Apr 3-10: rebuttal period 🎉/💀 May 6: Author notification #UAI2025 #ML #stats #learning #reasoning #uncertainty

03815

Reposted by Daniel Csillag

Omar Rivasplata @omarrivasplata.bsky.social · 15/12/2024

📢 Outstanding PhD student wanted! 🤓 The successful candidate will be based at The University of Manchester Dept. of CS ✨ to work on learning theory and methods for novel types of distributional shifts, co-supervised with @samikaski.bsky.social ⏳ DL 31.Jan.2025. www.findaphd.com/phds/project...

findaphd.com

Learning theory and methods for novel types of distributional shifts. at The University of Manchester on FindAPhD.com

PhD Project - Learning theory and methods for novel types of distributional shifts. at The University of Manchester, listed on FindAPhD.com

094

Daniel Csillag @dccsillag.xyz · 13/12/2024

Everyone should know how to do binary search and bracketing

021

Reposted by Daniel Csillag

Michael Tran @huytransformer1.bsky.social · 20/11/2024

go.bsky.app/7U9DDTv

2236

Daniel Csillag @dccsillag.xyz · 23/11/2024

@johndcook.bsky.social I just read your recent 'Categorical Data Analysis' post www.johndcook.com/blog/2018/04.... As a ML&stats person who tried to incorporate category theory in my work, I'd say the problem is that it just doesn't look like CT adds much. Whenever I try to incorporate it, ..

johndcook.com

Categorical Data Analysis

Categorical data analysis could mean a couple different things. One is analyzing data that falls into unordered categories (e.g. red, green, and blue) rather than numerical values (e.g. height in cent...

110

Reposted by Daniel Csillag

Towaki Takikawa @yongyuanxi.bsky.social · 21/11/2024

Many think Rust is primarily about memory safety, but the real reason to use Rust is developer productivity (for C++-like system programming) from having a sane package manager and type system.

2946225

Reposted by Daniel Csillag

Martin Trapp @trappmartin.eurosky.social · 20/11/2024

For those who don’t know yet, I am organising an online talk series together with Arno Solin on “Advances in Probabilistic Machine Learning (APML)”. It’s free for everyone to join and support early career researchers! You can register and check out the schedule here: aaltoml.github.io/apml/

aaltoml.github.io

Seminar on Advances in Probabilistic Machine Learning

This seminar series aims to provide a platform for young researchers (PhD student or post-doc level) to give invited talks about their research, intending to have a diverse set of talks & speakers on ...

29329

Daniel Csillag @dccsillag.xyz · 21/11/2024

Anthropic published a paper telling people to use confidence intervals for evals. I now await for their next paper, which will explain multiple comparisons to the LLM people

110

Reposted by Daniel Csillag

arxiv stat.ML @arxiv-stat-ml.bsky.social · 05/11/2024

Daniel Csillag, Claudio Jos\'e Struchiner, Guilherme Tegoni Goedert Strategic Conformal Prediction arxiv.org/abs/2411.01596

011

Daniel Csillag @dccsillag.xyz · 27/10/2024

NeurIPS registrations following a *randomized lottery* is absurd to me. Especially once you consider all the logistics involved with attendance (e.g., flights, stay, etc.)

000

Reposted by Daniel Csillag

Rafael Izbicki @rafael-izbicki.bsky.social · 21/09/2024

Ran my text through ChatGPT arguing why oversampling usually doesn’t make sense, and it flipped my argument to support oversampling instead. A reminder that AI often defaults to consensus—even when it's wrong. Always double-check! #AI #MachineLearning #Statistics

071

Daniel Csillag @dccsillag.xyz · 10/09/2024

tqdm really feels like the best library ever written in the history of ever. It's so good...

100

Reposted by Daniel Csillag

Hecate @ohmygoedel.bsky.social · 14/06/2023

A thread about some of my favorite books: 0. Logic and Structure. Dirk van Dalen.

2196

Daniel Csillag @dccsillag.xyz · 04/09/2024

Gradient descent with random initialization avoids troublesome stationary points almost surely. #🧮

Let $f : RR^n -> RR$ be twice continuously differentiable and assume its gradient is Lipschitz.
Consider gradient descent with random initialization:

$ x_0 ~ nn(bold(0),I), wide wide x_(t+1) = x_t - eta nabla f(x_t). $

If the procedure converges to some point $macron(x)$ and $eta$ is sufficiently small, then, almost surely:

(i) It is a critical/stationary point: $nabla f(macron(x)) = 0$.
(ii) It is a second-order stationary point: the Hessian $nabla^2 f(macron(x))$ is positive-semidefinite.

This means that it will not converge to maxima, nor most other spurious stationary points.

So, even though GD only uses first-order information (the gradient), we get a second-order guarantee "for free!"

170

Daniel Csillag @dccsillag.xyz · 02/09/2024

Useful fact: you can bound the gap between any two probability distributions using f-divergences. A rather nice case is with the Pearson chi-squared f-divergence.

000

Daniel Csillag @dccsillag.xyz · 01/09/2024

Just arrived here! Liking bsky so far, we'll see how it goes. Follow for occasional posts on math, ML and stats :)

110