I’m looking to collaborate with an(some) enthusiastic hacker(s) on a prototype / PoC application of HTM. For instance, perhaps a cat/dog tracking app could be something fun to get started with, but I’m open to brainstorm other ideas. This would be a side project mainly on the weekend and evenings. If you’re in the Bay Area, that’d be a big plus, however, I won’t preclude an internet collab if the time zones work reasonably well. And, while I’m open to all sorts of levels of experience with HTM, you should have some basic programming experience for this to productive. So if this sounds enticing don’t hesitant to contact me. Cheers!
Hi,
I can work remote ( from Turkey ).
I am interested but I don’t have experience in HTM but I have experince in programming ( java, javascript etc… )
I am studying HTM to see possibility to implement it on elasticsearch,
actually I want to build an anomally detection using HTM on top of elasticsearch ( if possible )
Is there anything I can do to help facilitate your project? You all can create a private message thread on the forum that includes all interested parties to have more private conversations, or you could continue posting your ideas for projects here on #htm-hackers. Let me know if you need something.
I’d love to get in on this when I’m finished launching my Kickstarter campaign. It should be concluded near the end middle of July.
I’ve actually had a pet project on my mind, because my son has two cats. I love the little guys, but the one regularly meows and cries quite loud. I thought, maybe a good way to keep him quite is give him something to do. I’ve seen laser pointers at Fry’s which wiggle around. But, I thought I could do better than that. What if I attached a laser pointer to a robotic arm, and programmed and rewarded an algorithm if it kept the cat’s attention for longer periods of time. Or perhaps, during certain times of the day when he’s more likely to meow loudly.
In my mind, a project like that might require cat recognition, and also recognize when the cat’s playing vs just lying about in the camera’s field of view.
Anyway, seems like it would be a fun way to advance development. I’d love to be a part of it. I estimate I’d be more available after mid-July.
I’m interested. I think that there are tools and apps out there that let
you set up virtual hackathons. I think we used one for the last NuPIC
hackathon . Maybe we can self organize this.
-Valter
@rhyolight thanks for asking, I’m making this up as I go. I’ve been private messaging and skype calling people to have an initial conversation. Intuitively I think a slack channel would be helpful coordinating over disparate TZs. In fact I made one, and got busy, but it’s still there: https://htm-community.slack.com
Hi @brev! How are you? I spoke with Paul, Deric, & Mike in this thread who were working on some interesting projects. But then I ended up getting busy with an ML project in astronomy (https://github.com/lsst-epo/light_curve_ml) for the past several months. But now I’d really like to work on HTM again.
Hi Wilson and others, I am researching the environments provided by Unity’s ML agents and OpenAI’s gym right now. I’d like to know what are the inputs that can be sensed and actuators that can be controlled. One issue is that a lot of these environments are influenced by RL tasks/problems and Numenta hasn’t developed “HTM-RL” (yet). Seems like long-term these simulators/environments/games will be interesting test beds.
I’d like to stick close with Numenta’s recent and ongoing work on grid cells. So my initial thought is to build simple examples illustrating these nascent theories one of these simulators. We could do a 3D version of the oft-cited touching-a-coffee-cup example. This has already been done in the abstract, but we get the hands on experience and familiarity with a simulator.
I’ve been looking for an opportunity to collaborate on a hack (the recent hackathon I participated in at Ericsson got me hooked). Just let me know where you need input. My Unity skills aren’t the best, but I have some experience.
I was trying to implement an OpenAI CartPole client myself. I have some ideas on how to do this.
However, being new to Nupic coding, I’m not sure how to send the observation data into the network. All the example code uses csv file readers for input.
What’s the easiest way of doing this in Nupic?