Latest episode is out, my dear #Bayesians! A deep dive into #AmortizedInference, what it looks like in practice, and how to teach it to your AI agents. Tune if you wanna see how to do fast, amortized inference that scales -- live, demoed by Stefan 😉

Pierre-Simon Laplace@learnbayesstats.bsky.social · 3mo ago

🎙️ New episode alert! In this episode @alex-andorra.bsky.social & Stefan Radev dive into amortized inference, train a neural net once on sims, deploy on real data as many times as you want. They cover sim-to-real, psych & neuro as test beds, honest failure modes and more ... lnkd.in/dCY85k4g

🧠 Check out the classic examples from Bayesian Cognitive Modeling: A Practical Course (Lee & Wagenmakers, 2013), translated into step-by-step tutorials with BayesFlow! Interactive version: kucharssim.github.io/bayesflow-co... PDF: osf.io/preprints/ps...

Finite mixture models are useful when data comes from multiple latent processes. BayesFlow allows: • Approximating the joint posterior of model parameters and mixture indicators • Inferences for independent and dependent mixtures • Amortization for fast and accurate estimation 📄 Preprint 💻 Code

A reminder of our talk this Thursday (30th Jan), at 11am GMT. Paul Bürkner (TU Dortmund University), will talk about "Amortized Mixture and Multilevel Models". Sign up at listserv.csv.warwick... to receive the link.

4/ Amortized Bayesian Workflow (Extended Abstract) jointly led by @marvinschmitt.com and @chengkunli.bsky.social , and with @avehtari.bsky.social @paulbuerkner.com @stefanradev.bsky.social MCMC + amortized methods for the best of both worlds (speed & guarantees!) arxiv.org/abs/2409.04332

Amortized Bayesian Workflow (Extended Abstract)

Bayesian inference often faces a trade-off between computational speed and sampling accuracy. We propose an adaptive workflow that integrates rapid amortized inference with gold-standard MCMC techniqu...

arxiv.org

Any single analysis hides an iceberg of uncertainty. Sensitivity-aware amortized inference explores the iceberg: ⋅ Test alternative priors, likelihoods, and data perturbations ⋅ Deep ensembles flag misspecification issues ⋅ No model refits required during inference 🔗 openreview.net/forum?id=Kxt...

Prior specification is one of the hardest tasks in Bayesian modeling. In our new paper, we (Florence Bockting, @stefanradev.bsky.social and me) develop a method for expert prior elicitation using generative neural networks and simulation-based learning. arxiv.org/abs/2411.15826

Expert-elicitation method for non-parametric joint priors using normalizing flows

We propose an expert-elicitation method for learning non-parametric joint prior distributions using normalizing flows. Normalizing flows are a class of generative models that enable exact, single-step...

arxiv.org