Jan Boelts
@janboelts.bsky.social
Researcher at appliedAI Institute for Europe.
Working on simulation-based inference and responsible ML
Working on simulation-based inference and responsible ML
On the maintainer side, we gained valuable insights into where users struggle with API limitations or missing docs β fuel for future development @sbi-devs.bsky.social
Great team work with @danielged.bsky.social and collaboration
@tommoral.bsky.social, Pedro Rodriguez and many more
Now π π¨π πΏ
Great team work with @danielged.bsky.social and collaboration
@tommoral.bsky.social, Pedro Rodriguez and many more
Now π π¨π πΏ
January 23, 2026 at 6:51 PM
On the maintainer side, we gained valuable insights into where users struggle with API limitations or missing docs β fuel for future development @sbi-devs.bsky.social
Great team work with @danielged.bsky.social and collaboration
@tommoral.bsky.social, Pedro Rodriguez and many more
Now π π¨π πΏ
Great team work with @danielged.bsky.social and collaboration
@tommoral.bsky.social, Pedro Rodriguez and many more
Now π π¨π πΏ
Projects spanned epidemiology, planetary physics, oceanography, gravitational waves, marine weather routing, particle physics, and medical imaging. π΅βπ«
Always amazing to see the range of areas where SBI can help.
Always amazing to see the range of areas where SBI can help.
January 23, 2026 at 6:51 PM
Projects spanned epidemiology, planetary physics, oceanography, gravitational waves, marine weather routing, particle physics, and medical imaging. π΅βπ«
Always amazing to see the range of areas where SBI can help.
Always amazing to see the range of areas where SBI can help.
SBI Hackathon Grenoble is a wrap! π
35 researchers and a great hybrid format of 1.5 days of tutorials + 1.5 days of applied hackathon. Many went from βhaving heard of sbiβ to applying full SBI workflows to their own research projects.
π§΅π
35 researchers and a great hybrid format of 1.5 days of tutorials + 1.5 days of applied hackathon. Many went from βhaving heard of sbiβ to applying full SBI workflows to their own research projects.
π§΅π
January 23, 2026 at 6:51 PM
SBI Hackathon Grenoble is a wrap! π
35 researchers and a great hybrid format of 1.5 days of tutorials + 1.5 days of applied hackathon. Many went from βhaving heard of sbiβ to applying full SBI workflows to their own research projects.
π§΅π
35 researchers and a great hybrid format of 1.5 days of tutorials + 1.5 days of applied hackathon. Many went from βhaving heard of sbiβ to applying full SBI workflows to their own research projects.
π§΅π
On my way from Munich to Grenoble π to co-lead a 3-day SBI tutorial + hackathon together with @danielged.bsky.social, organised by Pedro Rodriguez and @ugrenoblealpes.bsky.social.
Excited to meet researchers from across France, many bringing their own simulators π
Excited to meet researchers from across France, many bringing their own simulators π
January 20, 2026 at 7:31 PM
On my way from Munich to Grenoble π to co-lead a 3-day SBI tutorial + hackathon together with @danielged.bsky.social, organised by Pedro Rodriguez and @ugrenoblealpes.bsky.social.
Excited to meet researchers from across France, many bringing their own simulators π
Excited to meet researchers from across France, many bringing their own simulators π
After years of working on SBI methods and the sbi toolbox, we finally wrote the practical guide we wished had existed when we started.
Grateful to have collaborated with researchers across many institutions to consolidate what we've learned about making these methods work in practice!
Grateful to have collaborated with researchers across many institutions to consolidate what we've learned about making these methods work in practice!
Simulation-based inference (SBI) has transformed parameter inference across a wide range of domains. To help practitioners get started and make the most of these methods, we joined forces with researchers from many institutions and wrote a practical guide to SBI.
π Paper: arxiv.org/abs/2508.12939
π Paper: arxiv.org/abs/2508.12939
Simulation-Based Inference: A Practical Guide
A central challenge in many areas of science and engineering is to identify model parameters that are consistent with prior knowledge and empirical data. Bayesian inference offers a principled framewo...
arxiv.org
November 21, 2025 at 4:04 PM
After years of working on SBI methods and the sbi toolbox, we finally wrote the practical guide we wished had existed when we started.
Grateful to have collaborated with researchers across many institutions to consolidate what we've learned about making these methods work in practice!
Grateful to have collaborated with researchers across many institutions to consolidate what we've learned about making these methods work in practice!
Reposted by Jan Boelts
π sbi participated in GSoC 2025 through @numfocus.bsky.social and it was a great success: our two students contributed major new features and substantial internal improvements: 𧡠π
October 17, 2025 at 1:30 PM
π sbi participated in GSoC 2025 through @numfocus.bsky.social and it was a great success: our two students contributed major new features and substantial internal improvements: 𧡠π
Haha, yes! We are working on adding support for numpyro and pymc next.
September 18, 2025 at 4:51 PM
Haha, yes! We are working on adding support for numpyro and pymc next.
Kudos to @sethaxen.com for implementing the Pyro wrapper that makes this possible (shipped in sbi v0.25)!
And thanks to @juanitorduz.bsky.social sharing the cookie factory exampleβit's a great accessible example for hierarchical inference.
Everything runs in Colab π
And thanks to @juanitorduz.bsky.social sharing the cookie factory exampleβit's a great accessible example for hierarchical inference.
Everything runs in Colab π
September 18, 2025 at 2:38 PM
Kudos to @sethaxen.com for implementing the Pyro wrapper that makes this possible (shipped in sbi v0.25)!
And thanks to @juanitorduz.bsky.social sharing the cookie factory exampleβit's a great accessible example for hierarchical inference.
Everything runs in Colab π
And thanks to @juanitorduz.bsky.social sharing the cookie factory exampleβit's a great accessible example for hierarchical inference.
Everything runs in Colab π
Materials from my EuroSciPy talk "Pyro meets SBI" are now available: github.com/janfb/pyro-meets-sbi
I show how we can use @sbi-devs.bsky.social-trained neural likelihoods in pyro π₯
Check it out if you need hierarchical Bayesian inference but your simulator / model has no tractable likelihood.
I show how we can use @sbi-devs.bsky.social-trained neural likelihoods in pyro π₯
Check it out if you need hierarchical Bayesian inference but your simulator / model has no tractable likelihood.
September 18, 2025 at 2:38 PM
Materials from my EuroSciPy talk "Pyro meets SBI" are now available: github.com/janfb/pyro-meets-sbi
I show how we can use @sbi-devs.bsky.social-trained neural likelihoods in pyro π₯
Check it out if you need hierarchical Bayesian inference but your simulator / model has no tractable likelihood.
I show how we can use @sbi-devs.bsky.social-trained neural likelihoods in pyro π₯
Check it out if you need hierarchical Bayesian inference but your simulator / model has no tractable likelihood.
Reposted by Jan Boelts
From hackathon to release: sbi v0.25 is here! π
What happens when dozens of SBI researchers and practitioners collaborate for a week? New inference methods, new documentation, lots of new embedding networks, a bridge to pyro and a bridge between flow matching and score-based methods π€―
1/7 π§΅
What happens when dozens of SBI researchers and practitioners collaborate for a week? New inference methods, new documentation, lots of new embedding networks, a bridge to pyro and a bridge between flow matching and score-based methods π€―
1/7 π§΅
September 9, 2025 at 3:00 PM
From hackathon to release: sbi v0.25 is here! π
What happens when dozens of SBI researchers and practitioners collaborate for a week? New inference methods, new documentation, lots of new embedding networks, a bridge to pyro and a bridge between flow matching and score-based methods π€―
1/7 π§΅
What happens when dozens of SBI researchers and practitioners collaborate for a week? New inference methods, new documentation, lots of new embedding networks, a bridge to pyro and a bridge between flow matching and score-based methods π€―
1/7 π§΅
Fun read of their amazing contributions to the SBI hackathon! π₯
The SBI-Pyro bridge that @sethaxen.com built has a lot of potential I believe. I'll actually be presenting this work at @euroscipy.bsky.social this Wednesday - excited to share this with a broader audience.
euroscipy.org/talks/KCYYTF/
The SBI-Pyro bridge that @sethaxen.com built has a lot of potential I believe. I'll actually be presenting this work at @euroscipy.bsky.social this Wednesday - excited to share this with a broader audience.
euroscipy.org/talks/KCYYTF/
August 18, 2025 at 12:15 PM
Fun read of their amazing contributions to the SBI hackathon! π₯
The SBI-Pyro bridge that @sethaxen.com built has a lot of potential I believe. I'll actually be presenting this work at @euroscipy.bsky.social this Wednesday - excited to share this with a broader audience.
euroscipy.org/talks/KCYYTF/
The SBI-Pyro bridge that @sethaxen.com built has a lot of potential I believe. I'll actually be presenting this work at @euroscipy.bsky.social this Wednesday - excited to share this with a broader audience.
euroscipy.org/talks/KCYYTF/
Reposted by Jan Boelts
More great news from the SBI community! π
Two projects have been accepted for Google Summer of Code under the NumFOCUS umbrella, bringing new methods and general improvements to sbi. Big thanks to @numfocus.bsky.social, GSoC and our future contributors!
Two projects have been accepted for Google Summer of Code under the NumFOCUS umbrella, bringing new methods and general improvements to sbi. Big thanks to @numfocus.bsky.social, GSoC and our future contributors!
May 20, 2025 at 10:50 AM
More great news from the SBI community! π
Two projects have been accepted for Google Summer of Code under the NumFOCUS umbrella, bringing new methods and general improvements to sbi. Big thanks to @numfocus.bsky.social, GSoC and our future contributors!
Two projects have been accepted for Google Summer of Code under the NumFOCUS umbrella, bringing new methods and general improvements to sbi. Big thanks to @numfocus.bsky.social, GSoC and our future contributors!
Great reminder of this sunny and so productive week in March and how much I enjoy being part of this dedicated and lovely group of SBI contributors! π€
Big thanks to all participants & co-organizers! π
Big thanks to all participants & co-organizers! π
Great news! Our March SBI hackathon in TΓΌbingen was a huge success, with 40+ participants (30 onsite!). Expect significant updates soon: awesome new features & a revamped documentation you'll love! Huge thanks to our amazing SBI community! Release details coming soon. π₯ π
May 12, 2025 at 2:39 PM
Great reminder of this sunny and so productive week in March and how much I enjoy being part of this dedicated and lovely group of SBI contributors! π€
Big thanks to all participants & co-organizers! π
Big thanks to all participants & co-organizers! π
Reposted by Jan Boelts
π₯³Great news, our JOSS paper "sbi reloaded" has been accepted! π
This community lead by the fine folks of @sbi-devs.bsky.social is very welcoming and super fun to work with! I learn with every discussion I have.
paper: joss.theoj.org/papers/10.21...
review: github.com/openjournals...
This community lead by the fine folks of @sbi-devs.bsky.social is very welcoming and super fun to work with! I learn with every discussion I have.
paper: joss.theoj.org/papers/10.21...
review: github.com/openjournals...
[REVIEW]: sbi reloaded: a toolkit for simulation-based inference workflows Β· Issue #7754 Β· openjournals/joss-reviews
Submitting author: @janfb (Jan Boelts) Repository: https://github.com/sbi-dev/sbi Branch with paper.md (empty if default branch): joss-submission-2024 Version: v0.24.0 Editor: @boisgera Reviewers: ...
github.com
April 11, 2025 at 7:16 AM
π₯³Great news, our JOSS paper "sbi reloaded" has been accepted! π
This community lead by the fine folks of @sbi-devs.bsky.social is very welcoming and super fun to work with! I learn with every discussion I have.
paper: joss.theoj.org/papers/10.21...
review: github.com/openjournals...
This community lead by the fine folks of @sbi-devs.bsky.social is very welcoming and super fun to work with! I learn with every discussion I have.
paper: joss.theoj.org/papers/10.21...
review: github.com/openjournals...
We have been thinking about this for a while and now itβs here πLooking forward to all the exciting SBI applications we will be discussing, and to onboarding new contributors! π
π Exciting news! We are lauching an sbi office hour!
Join the sbi developers Thursdays 09:45-10:15am CET via Zoom (link: sbi Discord's "office hours" channel).
Get guidance on contributing, explore sbi for your research, or troubleshoot issues. Come chat with us! π€
github.com/sbi-dev/sbi/...
Join the sbi developers Thursdays 09:45-10:15am CET via Zoom (link: sbi Discord's "office hours" channel).
Get guidance on contributing, explore sbi for your research, or troubleshoot issues. Come chat with us! π€
github.com/sbi-dev/sbi/...
April 4, 2025 at 8:31 AM
We have been thinking about this for a while and now itβs here πLooking forward to all the exciting SBI applications we will be discussing, and to onboarding new contributors! π
Reposted by Jan Boelts
It's been a blast, thanks to @sbi-devs.bsky.social ! This week's hackathon was phenomenal! π π The sbi hackathon welcomed about 25 people in TΓΌbingen with contributions spanning the globe , e.g. πΊπΈπ―π΅π§πͺπ©πͺ. Wanna see, what we did? Check out the PRsπ
github.com/sbi-dev/sbi/...
github.com/sbi-dev/sbi/...
Pull requests Β· sbi-dev/sbi
sbi is a Python package for simulation-based inference, designed to meet the needs of both researchers and practitioners. Whether you need fine-grained control or an easy-to-use interface, sbi has ...
github.com
March 21, 2025 at 8:45 AM
It's been a blast, thanks to @sbi-devs.bsky.social ! This week's hackathon was phenomenal! π π The sbi hackathon welcomed about 25 people in TΓΌbingen with contributions spanning the globe , e.g. πΊπΈπ―π΅π§πͺπ©πͺ. Wanna see, what we did? Check out the PRsπ
github.com/sbi-dev/sbi/...
github.com/sbi-dev/sbi/...
On my way to the SBI hackathon in TΓΌbingenβon a EuroCity thatβs overbooked, delayed, and mysteriously missing all reservation signs. People pacing the aisles, luggage blocking exits, a baby wailing in the distanceβ¦ ππ₯π΅βπ«
If only the German railway system were as user-friendly as the SBI package! π₯²
If only the German railway system were as user-friendly as the SBI package! π₯²
sbi 0.24.0 is out! π This comes with important new features:
- π― Score-based i.i.d sampling
- π Simultaneous estimation of multiple discrete and continuous parameters or data.
- π: mini-sbibm for quick benchmarking.
Just in time for our 1-week SBI hackathon starting tomorrow---stay tuned for more!
- π― Score-based i.i.d sampling
- π Simultaneous estimation of multiple discrete and continuous parameters or data.
- π: mini-sbibm for quick benchmarking.
Just in time for our 1-week SBI hackathon starting tomorrow---stay tuned for more!
March 16, 2025 at 1:33 PM
On my way to the SBI hackathon in TΓΌbingenβon a EuroCity thatβs overbooked, delayed, and mysteriously missing all reservation signs. People pacing the aisles, luggage blocking exits, a baby wailing in the distanceβ¦ ππ₯π΅βπ«
If only the German railway system were as user-friendly as the SBI package! π₯²
If only the German railway system were as user-friendly as the SBI package! π₯²
thanks for the response! yes, that makes sense. I was wondering how exactly you sample from the augmented posterior as this was not clear to me from reading the paper. E.g., in the SBI case (DDM example), how do you sample p(theta | x_i, x_o) given that the conditions are not iid?
February 11, 2025 at 3:16 PM
thanks for the response! yes, that makes sense. I was wondering how exactly you sample from the augmented posterior as this was not clear to me from reading the paper. E.g., in the SBI case (DDM example), how do you sample p(theta | x_i, x_o) given that the conditions are not iid?
Really neat idea!
I had a quick look and did not quite understand how the βaugmentedβ posterior is obtained or how you can sample from it to calculate the conditional ranks.
Thanks π
I had a quick look and did not quite understand how the βaugmentedβ posterior is obtained or how you can sample from it to calculate the conditional ranks.
Thanks π
February 6, 2025 at 7:37 PM
Really neat idea!
I had a quick look and did not quite understand how the βaugmentedβ posterior is obtained or how you can sample from it to calculate the conditional ranks.
Thanks π
I had a quick look and did not quite understand how the βaugmentedβ posterior is obtained or how you can sample from it to calculate the conditional ranks.
Thanks π
Actually, I think this can be useful for mixed-effects models as well: you define a hierarchical prior and simulator with your fixed effects model and then train NLE on single trial data. At inference time, you then condition on x and conditions for each subject, or for multiple subjects at once.
January 3, 2025 at 1:37 PM
Actually, I think this can be useful for mixed-effects models as well: you define a hierarchical prior and simulator with your fixed effects model and then train NLE on single trial data. At inference time, you then condition on x and conditions for each subject, or for multiple subjects at once.
Good question, thanks! When you train NLE on single trials, you can now run inference given a set of iid trials and additionally condition on corresp. set of varying experimental condition (without retraining). Perm. inv. nets are used in NPE settings, which usually requires retraining. Clarified?
January 2, 2025 at 3:07 PM
Good question, thanks! When you train NLE on single trials, you can now run inference given a set of iid trials and additionally condition on corresp. set of varying experimental condition (without retraining). Perm. inv. nets are used in NPE settings, which usually requires retraining. Clarified?
For everyone working with trial-based i.i.d. data and varying experimental conditions - we have you covered now!
You need to train NLE only once and then can run MCMC with multiple subjects, trials and conditions, etc.
Example: sbi-dev.github.io/sbi/dev/tuto...
Reach out on GitHub for questions πββοΈ
You need to train NLE only once and then can run MCMC with multiple subjects, trials and conditions, etc.
Example: sbi-dev.github.io/sbi/dev/tuto...
Reach out on GitHub for questions πββοΈ
January 2, 2025 at 9:18 AM
For everyone working with trial-based i.i.d. data and varying experimental conditions - we have you covered now!
You need to train NLE only once and then can run MCMC with multiple subjects, trials and conditions, etc.
Example: sbi-dev.github.io/sbi/dev/tuto...
Reach out on GitHub for questions πββοΈ
You need to train NLE only once and then can run MCMC with multiple subjects, trials and conditions, etc.
Example: sbi-dev.github.io/sbi/dev/tuto...
Reach out on GitHub for questions πββοΈ
Would you mind adding me as well? I work on SBI and related methods. Thanks!
December 2, 2024 at 10:33 AM
Would you mind adding me as well? I work on SBI and related methods. Thanks!
So happy to be part of this project and see it growing!
Many thanks to all contributors and users for making it possible π
Many thanks to all contributors and users for making it possible π
The sbi package is growing into a community project π To reflect this and the many algorithms, neural nets, and diagnostics that have been added since its initial release, we have written a new software paper π Check it out, and reach out if you want to get involved: arxiv.org/abs/2411.17337
sbi reloaded: a toolkit for simulation-based inference workflows
Scientists and engineers use simulators to model empirically observed phenomena. However, tuning the parameters of a simulator to ensure its outputs match observed data presents a significant challeng...
arxiv.org
November 27, 2024 at 1:59 PM
So happy to be part of this project and see it growing!
Many thanks to all contributors and users for making it possible π
Many thanks to all contributors and users for making it possible π
Reposted by Jan Boelts
Hello, world! We are a community-developed toolkit that performs Bayesian inference for simulators. We support a broad range of methods (NPE, NLE, NRE, amortized and sequential), neural network architectures (flows, diffusion models), samplers, and diagnostics. Join us!
November 18, 2024 at 7:26 AM
Hello, world! We are a community-developed toolkit that performs Bayesian inference for simulators. We support a broad range of methods (NPE, NLE, NRE, amortized and sequential), neural network architectures (flows, diffusion models), samplers, and diagnostics. Join us!