Education, Expectation-Maximisation, Evolution

hello everyone and welcome back to active inference insights as always I'm your host Daris p uh py Wayne can't even say my own name anymore and today I'm joined by Dr Sam sanjie namjoshi sanjie is an expert in data science computer vision machine learning and bioinformatics with strong teaching experience in bioinformatics Biochemistry and computational Neuroscience he is also what I would say is a World expert in Act of inference uh in fact he's currently putting together a textbook which explains the fundamentals of the theory which will act as a wonderful addition to the education resources we as a community already have sorry for butchering that intro but sanjie welcome to the show uh we've got past my we've got past my least favorite part of the show which is the introduction so we're here thank you again for for joining me yeah thank you so much for inviting me really appreciate it no no it's it's it's really is a pleasure because I have been you know it's the least I can do I've been benefiting a lot from your textbook um which is not out so people if they're if they're sort of scrambling on on the internet to try and find your textbook they may only find the parpet Zulo and friston one um so you may want to start just by saying sort of well I think today's episode is going to be slightly different um normally we sort of do some deep dives into people's papers but I think today we're going to start at least by talking about the kind of education side of Act of inference um so nothing too technical necessarily but just s of thinking about how these ideas are permeating academic culture intellectual culture and then sort of broader culture but maybe where we could start I sort of gave a very brief introduction there is with you just kind of outlining how you found active inference uh because people might not be aware of your work sure yeah so the my path to active inference um was really unusual I think what was really unique about it was I had being at a point in my career where I was starting to figure out where all my interests were but I didn't know that active inference actually existed so that was the journey toward that was that moment of finding out that all of my research interests were colliding in one place and if we go back really far to just the beginning um when I was doing my undergraduate research I was working in neurobiology I also worked in genetics um I did a lot of work in Biochemistry and a lot of my my PhD work was actually molecular Neuroscience but I had wanted to move into more technical Direction so I did do uh my informatics work in my PhD uh which introduced me more to some of the techniques and skills that I would need um in computational Neuroscience which is when my postto first post was in in that area and it was at the end of my postto when I was getting close to leaving Academia that I discovered active inference while just browsing papers writing a grant um I was the work I was doing was uh modeling human behavior with reinforcement learning so there was definitely an overlap in relationship ship active inference but um my adviser has said yeah yeah we we know little about active inference we don't really use it in the lab but um it's something we're interested in and so that kind of just led me in that direction and reading some of Carl's original papers uh that just got me off on that path and I became really excited about learning more about it and that's kind of where I stumbled into the original problems of there's just no resources out there this was around 2018 yeah 2018 there were definitely no resources out there uh we're doing a bit better now we have the Institute we have the textbook by par psul and Frist which people can buy and then hopefully your one which is coming out um so want to go back it's it's interesting you said sort of the skills that you learned in doing bioinformatics were you know really contributed to your capacity to sort of pick up active inference quite…