Out now in TMLR! 🚀
We formalize how epistemic priors transform Expected Free Energy minimization into a standard variational objective.
This allows us to frame planning as a continuous variational optimization problem, moving away from combinatorial tree search. 👇
The result: a principled framework for embodied, agentic AI — robots, drones, and autonomous systems that perceive, learn, and act on-the-fly in real time.
I recently had the pleasure to lecture at the Machine Learning Summer School in Melbourne https://t.co/rEjkuAYqAZ on Bayesian Machine Learning → Active Inference.
All materials (slides + notebooks) available at https://t.co/97ftJyVqxb . Thread below 👇
Finally: Active Inference (AIF). AIF extends BML to embodied agents with a full commitment to variational inference for state estimation, learning, planning, and control. See https://t.co/74p1NV1QCa
This is what we ask ourselves in our latest paper (https://t.co/nH8BvgdvAp), where we solve Minigrid environments with Active Inference. (Spoiler, it works!)
Code available at https://t.co/3W3BzTaPkX
Done together with @mlmykola , @ThijsvdLaar and @bertdv0
What's even nicer: because our method injects priors locally, everything still works within @ReactiveBayes ' RxInfer.jl using message passing. Special thanks to my colleagues at @LazyDynamics for making this possible!
In Active Inference, a lot of time is spent on computing Expected Free Energy. What if we could tweak the generative model such that EFE can be minimised with traditional variational inference methods?
Backed Trojan Robotics (Team 24090) at the FIRST® Tech Challenge European Premier Event in Eindhoven (July 1–5) . They hustled—coding, building, troubleshooting—and came away with 3rd in the Think Award. Proud to support their next steps. 🚀 #FTC#Robotics#STEM
@MovingFramesP@fchollet Technically, FEP leads to simultaneously active inference (updating states), active learning and active model selection. The term “active inference” is usually interpreted as an umbrella over all active free energy minimizing processes.
Agreed with @fchollet on FEP (https://t.co/uiCiCLoLnM), but FEP is more than a pretty good idea, and there are more benefits to realizing an agent as an active inference (AIF) process beyond active data selection. I will mention a few below:
Some more refs: for implementation, https://t.co/rib0JP4Ks4 (toolbox), and https://t.co/uFI4ZLsdqD (AIF planning as inference), and https://t.co/9LqlAmLJTX (company). Also check out https://t.co/070D6VREic , https://t.co/Ez3FGUrtr8, and https://t.co/luV2XD7Cum .
Agreed with @fchollet on FEP (https://t.co/uiCiCLoLnM), but FEP is more than a pretty good idea, and there are more benefits to realizing an agent as an active inference (AIF) process beyond active data selection. I will mention a few below:
(6) Finally, FEP is more than a pretty good idea as it can be derived from first principles by information theory, see e.g., blog at https://t.co/IcPKWIGyTf plus refs. An AIF process avoids ad hoc design choices often found in man-made AI algorithms.