Ultra-efficient machine learning transistor cuts AI energy use by 99% (opens in new tab)

(newatlas.com)

28 points0xa22y ago9 comments

9 comments

You have to go from this article to the Northwestern press release to the source research paper to learn that this is for simple SVMs in wearable devices like for healthtech.

Neat, but it doesn't really have any bearing on the deep learning and generative AI trends.

heisgone2y ago

Are those transistor analog? It seems that AI would be fine with an analog weight evaluation and a lot of energy is wasted doing numerical calculations.

variadix2y ago

My understanding is that this doesn’t work for back propagation because it needs to know the forward pass computation exactly to adjust the weights. Analog computation introduces error in the form of noise (thermal, environmental, etc) which makes training on analog a non starter. It probably works for inference if the error is low enough, but training is where most of the power is spent unless it’s massively deployed.

heisgone2y ago

Thanks. That's interesting. If I understand correctly what you say, analog cannot be used for training but might be used to bake a trained static model. I think things are evolving too fast now for pre-trained models to be baked in silicon but it's a matter of time for it to become common.

1 more reply

tbalsam2y ago

You will get error regardless of whether it is analog or digital.

Digital trades bandwidth for error correction, but this does not mean it is better as many ML algorithms are noise-tolerant and would benefit greatly from the increased throughput.

Like everything, I believe there are simply just tradeoffs.

kridsdale32y ago

I have an analog weight evaluation neural net in my own skull. It runs fine on a few Snickers bars.

smusamashah2y ago

This seems very exciting. How long will it take to see the light of day and actually used in some form?

hulitu2y ago

> Ultra-efficient machine learning transistor cuts AI energy use by 99%

By using a diode they could cut the energy use by 99,999 %

j / k navigate · click thread line to collapse

9 comments

heyitsguay2y ago

You have to go from this article to the Northwestern press release to the source research paper to learn that this is for simple SVMs in wearable devices like for healthtech.

Neat, but it doesn't really have any bearing on the deep learning and generative AI trends.

heisgone2y ago

Are those transistor analog? It seems that AI would be fine with an analog weight evaluation and a lot of energy is wasted doing numerical calculations.

variadix2y ago

heisgone2y ago

1 more reply

tbalsam2y ago

You will get error regardless of whether it is analog or digital.

Digital trades bandwidth for error correction, but this does not mean it is better as many ML algorithms are noise-tolerant and would benefit greatly from the increased throughput.

Like everything, I believe there are simply just tradeoffs.

kridsdale32y ago

I have an analog weight evaluation neural net in my own skull. It runs fine on a few Snickers bars.

smusamashah2y ago

This seems very exciting. How long will it take to see the light of day and actually used in some form?

hulitu2y ago

> Ultra-efficient machine learning transistor cuts AI energy use by 99%

By using a diode they could cut the energy use by 99,999 %

j / k navigate · click thread line to collapse