You're reading the public-facing archive of the Category Theory Zulip server.
To join the server you need an invite. Anybody can get an invite by contacting Matteo Capucci at name dot surname at gmail dot com.
For all things related to this archive refer to the same person.
Sean Gloumeau (Oct 11 2021 at 21:55):Hi, I'm currently taking a master in embedded systems and was curious to find some papers with applications of CT to communication systems, control systems, machine learning, robotics or any other topic that might be related to embedded systems (although the 4 listed are my main focus in the broad field). I hope to use the previous research to guide me in some direction for applying CT in my master's thesis next year!
Dylan Braithwaite (Oct 11 2021 at 22:48):There's been recent work proposing categories of optics as a language for describing "cybernetic systems". This describes the general theory: https://arxiv.org/abs/2105.06332 . The most developed examples are in game theory and machine learning, but the paper describes general systems with feedback/control. And I understand there's work going on to broaden the scope of the examples this is applied to.
There was also a recent survey paper with more general examples of CT in machine learning that might be a useful reference: https://arxiv.org/abs/2106.07032
There was a workshop earlier this year on category theory in robotics as well. I don't know much more about that, but it looks like there's some videos for the talks, or it might be worthwhile looking into the speakers' work: https://idsc.ethz.ch/research-frazzoli/workshops/compositional-robotics.html
John Baez (Oct 11 2021 at 22:51):@Georgios Bakirtzis works on categories and robotics and may have something to say.
John Baez (Oct 11 2021 at 22:52):But for starters you should browse some of his papers.
John Baez (Oct 11 2021 at 22:53):His papers with Vasilakopoulou, at least, use category theory.
Georgios Bakirtzis (Oct 11 2021 at 23:52):My dissertation probably gives the most comprehensive view (which subsumed the work I do with Vasilakopoulou and genovese)
Georgios Bakirtzis (Oct 11 2021 at 23:55):These days I am trying to do a couple of things with ct, mostly in the space of autonomous systems which are modleed with markov decision processes but I haven’t found any luck finding collaborators with time from the ct community. Language models from Oxford also seem to me like a great way to do requirement decomposition but again I haven’t found anybody that’s interested. I don’t like to pretend like I am an expert on CT so until this happens I am doing more traditional robotics stuff but I am always willing to talk to people about my ideas reasoning and experience in this space
Georgios Bakirtzis (Oct 11 2021 at 23:57):Another interesting direction would be porting some control synthesis stuff to cat lab which I also would do if I get a position after this one with students. It would be a great masters project in my book but you also have to speak to your advisor
Georgios Bakirtzis (Oct 12 2021 at 00:05):Anyways this is a paper introducing the cyber-physical systems problems to ACT people https://arxiv.org/pdf/2101.10484.pdf
This is the paper that shows how we can do stuff with ACT for cyber-physical people https://dl.acm.org/doi/abs/10.1145/3461669
Here is a paper introducing ACT to the cyber-physical systems security problem https://arxiv.org/abs/2103.00044 (we are reworking a major revision that should be out the next month)
Here is a paper talking to a broader audience (IEEE Computer) about ACT and systems modeling with Subrahmanian from CMU and NIST https://arxiv.org/abs/2105.12911
And here is my dissertation that has some more motivational stuff https://libraetd.lib.virginia.edu/public_view/cr56n179w
We also have a general project for this stuff here https://bakirtzis.net/talon/ and code here https://github.com/rgCategory/composition_notebook that I plan on cleaning up and making tutorial jupyter notebooks
You should definitely look at what @Gioele Zardini is doing as well in codesign for embodied intelligence and CT
Georgios Bakirtzis (Oct 12 2021 at 00:06):Thank you @John Baez for tagging me on this
Jules Hedges (Oct 12 2021 at 13:18):Arguably, the best way to find out what people are working is not to look for papers, but to watch the recorded talks from the last 2 or 3 years of the ACT conference
Georgios Bakirtzis (Oct 12 2021 at 13:31):To some, the project website has one of my talks on this stuff that I think is pretty good to get a sense of what I am doing/trying to do.
Sean Gloumeau (Oct 13 2021 at 10:34):These are great suggestions, thank you everyone I really appreciate it! I'll have a look at all the linked sites and ACT then