4 ms·

Wow. Deep learning would certainly not be my technique of choice on constrained architectures, but there are situations where you don't really have viable alter

Wow. Deep learning would certainly not be my technique of choice on constrained architectures, but there are situations where you don't really have viable alternatives right now, so I'm glad to see that's actually doable.

jononor 8y ago
What would be your goto techniques? Curious since I'm researching machine learning on microcontrollers.
- ianhowson 8y ago
  Decision trees, random forests, logistic regression and most of the boring old statistical classifiers work on anything down to an 8-bit micro with <1k RAM. SVMs are highly effective and don't need much more RAM than that if you're careful.
  - jononor 8y ago
    Yeah, that was my take as well. I started by implementing Random Forests, really fast and compact for even the smallest of microcontrollers. Will probably add some variant of boosting trees in the future. https://github.com/jonnor/emtrees https://github.com/jonnor/emtrees
rbanffy 8y ago
It really depends on what you call "constrained". For about £5 you can get a Linux-powered RISC machine with 512MB of RAM, a GPU, and rich IO capabilities. I have worked on large multi-user environments smaller than that powering dozens of serial terminals on everyone's desktops. That's a lot of compute power. What I wouldn't like to do is to run the training part on such small devices. If there is a good way to do incremental learning after you trained your model so it could continuously fine tune itself using the embedded hardware on a reasonable power budget, I'd go for it. And while you won't run large networks, you can probably get away with many smaller, more specialized ones.
- jononor 8y ago
  Price for compute is less and less constraining every year. However if running on battery the energy budget can be severely constraining. Also people just end up wanting to do more. Real-time video at decent framerate is still challenging for sub-100 USD devices. When that's easy, time for real-time 3d data (LIDAR etc)

jononor 8y ago

What would be your goto techniques? Curious since I'm researching machine learning on microcontrollers.

ianhowson 8y ago
Decision trees, random forests, logistic regression and most of the boring old statistical classifiers work on anything down to an 8-bit micro with <1k RAM. SVMs are highly effective and don't need much more RAM than that if you're careful.
- jononor 8y ago
  Yeah, that was my take as well. I started by implementing Random Forests, really fast and compact for even the smallest of microcontrollers. Will probably add some variant of boosting trees in the future. https://github.com/jonnor/emtrees https://github.com/jonnor/emtrees

rbanffy 8y ago

It really depends on what you call "constrained". For about £5 you can get a Linux-powered RISC machine with 512MB of RAM, a GPU, and rich IO capabilities. I have worked on large multi-user environments smaller than that powering dozens of serial terminals on everyone's desktops. That's a lot of compute power. What I wouldn't like to do is to run the training part on such small devices. If there is a good way to do incremental learning after you trained your model so it could continuously fine tune itself using the embedded hardware on a reasonable power budget, I'd go for it. And while you won't run large networks, you can probably get away with many smaller, more specialized ones.

jononor 8y ago
Price for compute is less and less constraining every year. However if running on battery the energy budget can be severely constraining. Also people just end up wanting to do more. Real-time video at decent framerate is still challenging for sub-100 USD devices. When that's easy, time for real-time 3d data (LIDAR etc)