4 ms·
Haven't looked into the paper. However you don't need a NN to simulate painting. Basically it's a search problem. See my results here: https://forwardscatteri
by Ono-Sendai 9y ago
Haven't looked into the paper.
However you don't need a NN to simulate painting. Basically it's a search problem.
See my results here: https://forwardscattering.org/post/42 https://forwardscattering.org/post/42
- nthngnss 9y agoWhile it is true that one can simulate painting by search (in some sense), the problem is that (naive) search doesn't always work (we have an example in the paper). Moreover, training an agent has a benefit of fast test-time inference (i.e., you give an image, and it paints it almost instantly). Of course, we haven't achieved the ultimate goal yet but it's a step in that direction.
- Ono-Sendai 9y agoThanks for the reply. Where in the paper is the example given? Skimmed it but didn't see it. (Edit: nevermind, see it now) Fast painting is a benefit I guess. My search/painting program is very computationally intensive. Edit: I think I see the point of the paper now. Unguided search is going to be difficult in high-dimensional search spaces like this. So the NNs become a hopefully-effective heuristic guiding the search.
- nthngnss 9y agoYup, exactly! Btw, didn't realize those were your results (just mindlessly clicked on the link). Very cool stuff!
- Ono-Sendai 9y agoThanks. The (semi-) obvious next step is to do object/digit recognition with a Bayesian probability calculation, with probabilities bases on this image reconstruction process. In other words, we choose e.g. digits based on how likely they are to have been drawn to give the target image. I have experimented a little with this idea, but with no successful results so far (plain old NNs still beat it).
- cttet 9y agoThe main difference is that for search you need to specify an image to emulate it, while the RL method aim to generate new ones after observing some examples.