Sign in

Luigi Acerbi

@lacerbi.bsky.social
2.3K followers 219 following 256 posts

Assoc. Prof. of Machine & Human Intelligence | Univ. Helsinki - Finnish Centre for AI - ELLIS Institute Finland | Bayesian ML & probabilistic modeling | lacerbi.github.io

PostsRepliesMedia
Reposted by Luigi Acerbi
Weiji Ma @weijima.bsky.social · 30/09/2026
Anybody can make a science film now! Turn a dense technical paper into a video accessible to a broad audience. I am quite impressed. The paper might be my last on multisensory perception. This project was started in 2008 and took 18 years to get published. Read more: acerbilab.org/one-cause-or...
acerbilab.org
One cause or two?
How does your brain know that a sight and a sound belong together? A short film about causal inference in perception, and about the shapes of its two ingredients: the blur of our senses, and our expec...
093
Luigi Acerbi @lacerbi.bsky.social · 30/09/2026
1/ A short film about how your brain solves the "one thing or two?" problem, based on our recent paper with Shuze Liu, Trevor Holland and @weijima.bsky.social in PLoS Comp Biol.
173
Luigi Acerbi @lacerbi.bsky.social · 07/09/2026
1/ Dear all, I am hiring a PhD student and/or postdoc to join my group at the @univhelsinkics.bsky.social and @ellisinstitute.fi! The general topic is Amortized Probabilistic Machine Learning for Decision-Making with applications in healthcare and scientific modelling. DL September 21 More info 👇
2139
Reposted by Luigi Acerbi
Alexandre Andorra @alex-andorra.bsky.social · 27/07/2026
Episode 161 of the show is out, with @lacerbi.bsky.social ! We dive into all things neural processes, fast #Bayesian inferences, and LLMs -- hope you enjoy 🖖
032
Luigi Acerbi @lacerbi.bsky.social · 17/07/2026
1/ Great chat with Alex Andorra aka @learnbayesstats.bsky.social about efficient inference, from amortized to surrogate-based approaches and a variety of related topics (prior-fitted networks, foundation models for inference and planning, etc.), many of which are neural processes in a trenchcoat.
2144
Luigi Acerbi @lacerbi.bsky.social · 18/06/2026
... and with the video. :) See thread for more info!
001
Luigi Acerbi @lacerbi.bsky.social · 18/06/2026
1/ Amortized probabilistic methods like our amortized conditioning engine (ACE) can solve regression, simulator-based inference, optimization, active learning in a single forward pass... Try it in your browser & see all distributions change in real time! Links and more info👇
172
Reposted by Luigi Acerbi
ELLIS Institute Finland @ellisinstitute.fi · 02/06/2026
Building the world's first nationwide AI model for healthcare, with high-quality registry data + next-gen machine learning. ELLIS Institute Finland's health moonshot is lifting off, with a boost from Business Finland's funding. Read more: www.ellisinstitute.fi/a-world-firs... 🤝 @fimm-uh.bsky.social
ellisinstitute.fi
A world-first nationwide AI model will boost decision-making in healthcare | ELLIS Institute Finland
FINe-Health Foundry will use Finland’s national health databases as part of a ‘Swiss Army knife’ AI model for healthcare.
2125
Reposted by Luigi Acerbi
Gus @gusthema.bsky.social · 05/05/2026
Gemma 4 models are great, but what if they were even faster and keeping the same quality? we got you covered! Today we are releasing the Gemma 4 MTP models to help with that: blog.google/innovation-a...
blog.google
Accelerating Gemma 4: faster inference with multi-token prediction drafters
An overview of how Multi-Token Prediction (MTP) drafters are making Gemma 4 models up to 3x faster at inference.
214518
Luigi Acerbi @lacerbi.bsky.social · 30/04/2026
1/ Wouldn't it be nice if you could perform Bayesian inference *efficiently* but also *reliably*? Amortized inference offers the former, while MCMC is often presented as the "gold standard" for accuracy and reliability. Enter the Amortized Bayesian Workflow...
12610
Reposted by Luigi Acerbi
David Duvenaud @davidduvenaud.bsky.social · 28/04/2026
Announcing Talkie: a new, open-weight historical LLM! We trained and finetuned a 13B model on a newly-curated dataset of only pre-1930 data. Try it below! with @alecrad.bsky.social and @nicklevin01.bsky.social
25615
Luigi Acerbi @lacerbi.bsky.social · 23/04/2026
If you are at @iclr-conf.bsky.social, please go and have a chat with Yang Yang, @nasrullohloka.bsky.social or @conorhassan.bsky.social at our posters! (re-posted after fixing the time zone...)
042
Reposted by Luigi Acerbi
ELLIS Institute Finland @ellisinstitute.fi · 22/04/2026
Our researchers are at #ICLR2026 with the latest machine learning advances. Check them out April 23-27 in Rio de Janeiro and on our site ➡️ www.ellisinstitute.fi/ellis-instit... @iclr-conf.bsky.social @csaalto.bsky.social @univhelsinkics.bsky.social @ellis.eu @samikaski.bsky.social
Our latest results: ICLR + AISTATS 2026.
ELLIS Institute Finland's researchers are at some of the year's major AI and ML conferences, presenting a total of 26 papers and posters at ICLR and AISTATS.
See all the publications on our site.
151
Reposted by Luigi Acerbi
Tommy Rochussen @rochussen.bsky.social · 22/04/2026
If you're at ICLR and want to chat about amortised inference/neural processes or probabilistic ML more broadly, swing by my poster between 10:30-13:00 on Thursday (poster session 1) in Pavilion 3 at location #206. The poster is about NP-style learning in BNNs to find good priors.
041
Luigi Acerbi @lacerbi.bsky.social · 22/04/2026
1/ Another @iclr-conf.bsky.social paper thread! Do you want to make SOTA probabilistic predictions using transformers & your dataset is a *set* (not a sequence or time series), so you care about permutation invariance... but also efficiency? Keep reading, we have exactly what you need. 👇
1117
Luigi Acerbi @lacerbi.bsky.social · 20/04/2026
1/ One of the issues of fully amortized inference / pretrained simulator-based inference is that you are stuck with the "prior" training distribution. What if you change your mind after training? In PriorGuide, one of our papers at ICLR this week, we allow the prior to be changed at runtime!
1248
Luigi Acerbi @lacerbi.bsky.social · 01/04/2026
Big day ahead for @chengkunli.bsky.social -- no pressure! www.helsinki.fi/en/faculty-s... @univhelsinkics.bsky.social Featuring @upicchini.bsky.social as our honoured Opponent, and your truly as the Custos. Looking forward to the big event.
helsinki.fi
Chengkun Li defends his PhD thesis on Surrogate-based methods for efficient Bayesian posterior computation | Faculty of Science | University of Helsinki
On Wednesday the 1st of April 2026, M.Eng. Chengkun Li defends his PhD thesis on Surrogate-based methods for efficient Bayesian posterior computation. The thesis is related to research done in the Mac...
170
Reposted by Luigi Acerbi
Jakob Macke @jakhmack.bsky.social · 28/03/2026
We’re launching a new AI Methods & Software Hub in the ML cluster and are looking for a Director to build and lead it! Shape how ML/AI and software drive scientific discovery—in close collaboration with AI and domain scientists—within an extremely vibrant and collaborative ecosystem!
0106
Luigi Acerbi @lacerbi.bsky.social · 24/03/2026
1. Dear colleagues, I finally got my first extremely well-crafted, highly manipulative nerd-sniping email from a prospective student whose core was 100% AI-assisted and that felt weird and I almost fell for it. We are in for a fun ride, aren't we?
140
Reposted by Luigi Acerbi
NeurIPS Europe @neuripseurope.bsky.social · 23/03/2026
Today NeurIPS is announcing our official satellite event in Paris. After responding to the call from Ellis following the success of EurIPS in December, we are pleased to reach a new milestone by joining forces with the NeurIPS organizing committee for the 2026 edition.
18933
Luigi Acerbi @lacerbi.bsky.social · 20/03/2026
𝐷𝑖𝑓𝑓𝑒𝑟𝑒𝑛𝑡𝑙𝑦!
060
Luigi Acerbi @lacerbi.bsky.social · 18/03/2026
The fitting ending for a great research journey!
041
Reposted by Luigi Acerbi
BAMB! @bamb-school.bsky.social · 06/03/2026
24-Hour Final Call for BAMB! 2026 ⏳ Join us in Barcelona (July 12–23) and learn from our expert faculty: @meganakpeters.bsky.social @marcelomattar.bsky.social @khamascience.bsky.social @thecharleywu.bsky.social Apply now: www.bambschool.org
bambschool.org
BAMB! 2026 | Barcelona Summer School for Advanced Modeling of Behavior
Intensive training for experienced researchers in cognitive science, computational neuroscience and neuro-AI. Five interconnected modules, expert faculty, hands-on projects. July 12-23, 2026.
11310
Luigi Acerbi @lacerbi.bsky.social · 11/02/2026
Yes, this still happens! Before doing a manual check, I deploy agents to do the first check for me (literally a /doublecheck skill in Claude Code which I force to call automatically at the end of each task). This often catches issues, including once a totally fabricated MCMC analysis...
110
Luigi Acerbi @lacerbi.bsky.social · 06/02/2026
A bit of a delayed celebration, but happy that our three submitted papers were accepted at @iclr-conf.bsky.social 2026! This was a... complicated year for ICLR, but hopefully now we can focus on the science.
1154
Reposted by Luigi Acerbi
Ethan Mollick @emollick.bsky.social · 27/01/2026
This game was 100% designed, made, and tested by Claude Code with one prompt to "make a complete Sierra-style adventure game with EGA-like graphics and text parser, with 10-15 minutes of gameplay." I gave two prompts to play test the game & deploy it. Play: enchanted-lighthouse-game.netlify.app
5899
Luigi Acerbi @lacerbi.bsky.social · 14/01/2026
1/ Excited to share our new work published in Transactions on Machine Learning Research (TMLR), Stacking Variational Bayesian Monte Carlo (S-VBMC)!
Stacking Variational Bayesian Monte Carlo paper in TMLR.
1142
Reposted by Luigi Acerbi
Thomas Dietterich @tdietterich.bsky.social · 28/12/2025
I'd like to propose the following norm for peer review of papers. If a paper shows clear signs of LLM-generated errors that were not detected by the author, the paper should be immediately rejected. My reasoning: 1/ #ResearchIntegrity
411628
Luigi Acerbi @lacerbi.bsky.social · 26/12/2025
What an amazing Yule gift from @stefanradev.bsky.social & colleagues: a tour-de-force tutorial on diffusion models for simulator-based inference. This is one of the most comprehensive and useful review/tutorials I have ever seen -- a must read! Kudos to all the authors! arxiv.org/abs/2512.20685
arxiv.org
Diffusion Models in Simulation-Based Inference: A Tutorial Review
Diffusion models have recently emerged as powerful learners for simulation-based inference (SBI), enabling fast and accurate estimation of latent parameters from simulated and real data. Their score-b...
0353
Reposted by Luigi Acerbi
David Rügamer @davidruegamer.bsky.social · 06/12/2025
EurIPS was pretty fun! Thanks to @lacerbi.bsky.social , @conorhassan.bsky.social , and all organizers of the Amortized ProbML team for the great workshop and for inviting me as a speaker! 🙌 Also great to be able to present our spotlight paper on sparsity on the main stage 😍
1185
Reposted by Luigi Acerbi
Constantin Rothkopf @c-rothkopf.bsky.social · 04/12/2025
Tobias Niehues @tobnie.bsky.social is presenting our work with @dominikstrb.bsky.social 'Amortized Bayesian decision-making for inferring decision-making parameters from behavior' at the Amortized ProbML Workshop and the @ellis.eu UnConference. Please come by our poster!
0173
Luigi Acerbi @lacerbi.bsky.social · 03/12/2025
1/ Very happy with this spotlight paper at @neuripsconf.bsky.social where we continue our "Amortize Everything" agenda. After the Amortized Conditioning Engine (Chang et al., AISTATS, 2025) which amortizes all sorts of inference tasks, here with ALINE we amortize both inference & design.
192
Reposted by Luigi Acerbi
Guy Moss @gmoss13.bsky.social · 01/12/2025
I’m super excited to present our new work in #Eurips2025 and #Neurips2025! We developed FNOPE: a new simulation-based inference (SBI) method which excels at inferring function-valued parameters! Paper: openreview.net/forum?id=yB5... Code: github.com/mackelab/fnope (1/9)
openreview.net
FNOPE: Simulation-based inference on function spaces with Fourier...
Simulation-based inference (SBI) is an established approach for performing Bayesian inference on scientific simulators. SBI so far works best on low-dimensional parametric models. However, it is...
1205
Luigi Acerbi @lacerbi.bsky.social · 28/11/2025
Are there non-predatory for-profit publishers? This "thing" clearly never went through peer review despite what they say. Well, maybe it did go through "peer review": LLM reviewers reviewed an LLM-written paper, fair enough.
050
Luigi Acerbi @lacerbi.bsky.social · 27/11/2025
If you are at ELLIS UnConference on Dec 2 (pre @euripsconf.bsky.social ), join us at our Amortized ProbML workshop! Come for our great speakers: @paulbuerkner.com Yingzhen Li @davidruegamer.bsky.social @liza-semenova.bsky.social & poster session; stay for our hot takes on amortized probML!
Amortized ProbML workshop page snapshot
1186
Luigi Acerbi @lacerbi.bsky.social · 03/11/2025
Very interesting work on emerging object-binding representations in vision transformers by @kordinglab.bsky.social arxiv.org/abs/2510.24709 This might seem an oddly specific property but the good ol' binding problem reflects a fundamental primitive of cognition, epistemology, you name it.
arxiv.org
Does Object Binding Naturally Emerge in Large Pretrained Vision Transformers?
Object binding, the brain's ability to bind the many features that collectively represent an object into a coherent whole, is central to human cognition. It groups low-level perceptual features into h...
070
Luigi Acerbi @lacerbi.bsky.social · 26/10/2025
I answered in thread but it might be of interest. My current advice is to use the stderr of the median (bootstrapped), with a separate metric for reliability if needed. The latter is less standard - we have been reporting e.g. 80% perf quantile and its stderr (bootstrapped).
072
Reposted by Luigi Acerbi
Ethan Mollick @emollick.bsky.social · 26/10/2025
Claude, GPT-5, Gemini, and Kimi: "write me a horror story done entirely in the dedications to six books (you can give me the title and author of each book as well)" ChatGPT and Claude did pretty well in different ways. Kimi did the usual (sounds good but meaning falls apart).
3363
Luigi Acerbi @lacerbi.bsky.social · 17/10/2025
**Submission DL extended to Oct 25!** You are still in time to submit a short abstract on amortized (simulator-based) inference (SBI), neural processes, prior-fitted networks (PFNs), amortized experimental design, foundation models, and related applications. Call: sites.google.com/view/amortiz...
sites.google.com
Amortized ProbML 2025 - Call for Papers
Accepted papers will be presented as posters during the workshop poster session on 2nd December, 2025. Please be aware that this is not the same as the ELLIS UnConference poster session.
041
Luigi Acerbi @lacerbi.bsky.social · 14/10/2025
Reminder that Amortized inference Workshop submissions at the ELLIS Unconference are still open until **Oct 16, 2025**. Only a short abstract (½ page), so go ahead! Workshop: Dec 2, 2025, co-located with EurIPS. Website: sites.google.com/view/amortiz...
lnkd.in
LinkedIn
This link will take you to a page that’s not on LinkedIn
0133
Luigi Acerbi @lacerbi.bsky.social · 01/10/2025
Do you like to train neural networks to solve all your nasty probabilistic inference and sequential design problems? Do you love letter salads such as NPs, PFNs, NPE, SBI, BED? Then no place is better than the Amortized ProbML workshop we are organizing at #ELLIS UnConference.
Amortized ProbML:
Unlocking Probabilistic Machine Learning for Real-world Inference & Design (🇩🇰)
ELLIS UnConference Workshop
Copenhagen | 2 December, 2025
1216
Reposted by Luigi Acerbi
Alexander Doria @dorialexander.bsky.social · 27/09/2025
And new paper out: Pleias 1.0: the First Family of Language Models Trained on Fully Open Data How we train an open everything model on a new pretraining environment with releasable data (Common Corpus) with an open source framework (Nanotron from HuggingFace). www.sciencedirect.com/science/arti...
817951
Reposted by Luigi Acerbi
Liu, Shuze @liushuze.bsky.social · 19/08/2025
@gershbrain.bsky.social and I have a new paper in PLOS Comp Bio! We study how two cognitive constraints—action consideration set size & policy complexity—interact in context-dependent decision making, and how humans exploit their synergy to reduce behavioral suboptimality. osf.io/preprints/ps...
osf.io
OSF
12611
Reposted by Luigi Acerbi
Aki Vehtari @avehtari.bsky.social · 13/08/2025
Posterior predictive checking of binary, categorical and many ordinal models with bar graphs is useless. Even the simplest models without covariates usually have such intercept terms that category specific probabilities are learned perfectly. Can you guess which model, 1 or 2, is misspecifed? 1/4
Useless posterior predictive checking bar graphs for Models 1 and 2
1366
Luigi Acerbi @lacerbi.bsky.social · 27/07/2025
Remember to follow the official rebuttal guide.
A 26-panel infographic titled “A to Z of How to Write a NeurIPS Rebuttal,” with each panel representing a letter of the alphabet and a corresponding step or emotion. The top-left panels (A to D) show calm, professional cartoon characters with labels like “Acknowledge,” “Be concise,” “Clarify,” and “Dispute.” As the panels progress, the characters grow increasingly stressed, disheveled, and wild-eyed. Midway through, with panels like “Justify choices,” “Keep calm,” and “Obsess,” the illustrations become more chaotic and colorful. Toward the end (letters U to Z), the images turn black and white, depicting characters with frantic, exaggerated expressions, culminating in “Zombie.” The visual style shifts from clean and organized to frenetic and distorted, representing the mental unraveling of a researcher under rebuttal pressure.
1111
Luigi Acerbi @lacerbi.bsky.social · 17/07/2025
Superlative initiative! Massive kudos to the organisation team for getting this going in such a short time!
061
Reposted by Luigi Acerbi
K. Thor Jensen @kthorjensen.bsky.social · 28/06/2025
The singularity is awesome
Google search for “all you can eat buffet near me”
Al Overview
To find an all-you-can-eat buffet near you, you can use online search engines like Google or Yelp, specifying "all you can eat buffet near me". These searches will typically provide a list of nearby restaurants offering buffet-style dining, including their addresses, contact information, and customer reviews. You can also refine your
272207264332
Luigi Acerbi @lacerbi.bsky.social · 26/06/2025
New blog post! You can train a neural network to *just do things* -- such as *predict the optimum of a function*. But how do you get a big training dataset of "functions with known optima"? Read the blog post to find out! (link 👇)
Screenshot of blog post titled "You can just predict the optimum"
3214
Luigi Acerbi @lacerbi.bsky.social · 26/05/2025
That's a pity. I am not a fan of this place either due to the lack of interesting AI/ML discussions, and that's the only reason I still set foot in Mordor. A shiver goes through my spine at the thought that LinkedIn might be a viable alternative.
1100
Reposted by Luigi Acerbi
Tomer Ullman @tomerullman.bsky.social · 22/05/2025
sometimes you need to exorcise images from your mind
0101