Applied Active Inference Symposium 2025 part 1

Greetings, everyone. Welcome to the 5th Applied Active Inference Symposium. It's November 12th, 2025. Some brief introduction slides and then to Carl Friston for a keynote address and discussion. So welcome to the Active Inference Institute. We are a non-profit, volunteer-run organization dedicated to improving the rigor, accessibility, and applicability of the Active Inference Framework. Since 2021, we have hosted an annual Applied Active Inference Symposium with Carl as the first presenter back in 2021, which serves as an end-of-year celebration and convening for our ecosystem. At the link symposium.activeinference.institute, you can look at all the previous programs and proceedings. We're here to highlight applied and translational work in Active Inference across domains and from groups around the world. Meanwhile, we service opportunities and needs of our ecosystem, get feedback, and facilitate connections among participants. That's you. You can join the Institute Discord, donate to support the work, stay tuned via newsletter, and actively infer on. About the program, there are three days of programs, and the bulk of the proceedings will be live-streamed to YouTube at the three links that are available at the program. And you can watch it live and re-watch it at the same link. And there may be some adjustment at the fine scale with time zones and pre-recorded and live and scheduling and all fun kinds of things. But big picture, there's going to be three live streams that will consist the bulk of the program. In the participant registration, we asked if people wanted their responses to be public. And for those who did, we integrated their responses. So you can read the raw responses to people's questions that people were asked, like what would be interesting for them? What would be useful? How can Active Inference have a big impact in 2026? And we also used LLM to integrate the responses and do embeddings on the responses, all kinds of cool stuff. So you can check out that open source code. A big thank you to our team of co-organizers. Alexandra Mikhailova, Andrew Pachea, Bradley Licia, Christian Martin, Corey Slater, Daniel Friedman, Maria Garcia, Maria Luisa Iannico, Pablo FM, Sylvia Zhang, and Zora Ramanijad. It was a great and streamlined process with these co-organizers. So absolutely was a team effort. And I thank them all again. If there's some kind of session presenter or development that you'd like to see in next year's symposium, you would be welcome to provide that feedback or even to step up and be a co-organizer. It's a lot of fun. Email us if you're interested. Also, thank you to our sponsors. Our first sponsor is Navigating Uncertainty, a book bringing the science of active inference and the free energy principle to life through story, art, and human experience by Anna Pereira with a foreword by Carl Friston. And Anna will have a presentation later in the program. And also our sponsor for this symposium, Upcycle Club, an independent social initiative that actively supports a select group of noteworthy institutions and innovative projects dedicated to fostering research, education, and original thinking. So now, without further ado, to Carl's keynote. And I thought, how to introduce this keynote? And on this slide, I have Carl thinking of the 2012 paper, The History of the Future of the Bayesian Brain, where there's a quote, which was about the neuroimaging community, but I believe it applies to the active inference community today as well. And he wrote, in short, the future of imaging neuroscience is to refine models of the brain to minimize free energy, where the brain refines models of the world to minimize free energy. This endeavor itself minimizes free energy because our community is itself a self-organizing system. And so here, from the past to the predictive future, we have generative AI AI and all these different domains of application for active inference. So hopefully a fittingly aesthetic…