Training computer vision models on random noise instead of real images

Re:Training computer vision models on random noise instead o... by fennecfoxen on Thursday December 09, 2021 @11:44PM (#29506490)

If I'm reading it right, it seems to be saying that if you throw random noise at your image classifier, it's able to do produce an intermediate artifact that you can train more effectively when you throw hot dogs at it. It won't be quite as good as you get with the best models, but it's strikingly good for random noise. The paper puts it like so: "Although GANs need real images to learn the network parameters, we show in this paper that they introduce a structural prior useful to encode image properties without requiring any training."

https://openreview.net/pdf?id=RQUl8gZnN7O

[ Parent | on HN ]

Re:Training computer vision models on random noise instead o... by samkater on Friday December 10, 2021 @12:11AM (#29506638)

I’m not an expert, but my light reading of the paper makes it seem like it isn’t “random noise” in the colloquial sense, but more like “there are properties to the randomness even though the images look like noise.”

The abstract seems to say that, but again, I could be misinterpreting: “Our findings show that it is important for the noise to capture certain structural properties of real data but that good performance can be achieved even with processes that are far from realistic.”

[ Parent | on HN ]

Re:Training computer vision models on random noise instead o... by jswulff on Friday December 10, 2021 @02:01AM (#29507206)

That is correct -- by "noise", we don't meet pixel noise, but instead a stochastic process (in fact, different processes different properties which we compare in the paper) from which we can sample large amounts of varied training images.

[ Parent | on HN ]

Re:Training computer vision models on random noise instead o... by Retric on Friday December 10, 2021 @06:13AM (#29508591)

Except your not actually training on noise, your training on procedurally generated images which include random modifications. It’s just the summery is really bad at describing the process.

[ Parent | on HN ]