Hacker News: ExUtumno

New comment by ExUtumno in "MarkovJunior: Probabilistic PL based on pattern matching and constraint prop"

ExUtumno — Thu, 02 Jun 2022 13:41:53 +0000

Author here, I'll try to answer questions/comments in this post!

New comment by ExUtumno in "MarkovJunior: Probabilistic PL based on pattern matching and constraint prop"

ExUtumno — Thu, 02 Jun 2022 13:40:33 +0000

Thanks for reposting, I don't mind.

New comment by ExUtumno in "Show HN: MarkovJunior, PPL based on pattern matching and constraint propagation"

ExUtumno — Thu, 02 Jun 2022 13:40:09 +0000

Thanks! I would have probably never known about Markov algorithms if I were not taught them in 8th grade =)

New comment by ExUtumno in "Show HN: MarkovJunior, PPL based on pattern matching and constraint propagation"

ExUtumno — Thu, 02 Jun 2022 13:36:38 +0000

Thank you!

For MarkovJunior, the recent projects that were impactful the most were Imagegram by Guilherme S. Tows [1] and Daniel Ritchie's dissertation [2] about PPLs for procgen. I took quite a different approach from Ritchie's though.

[1] https://zaratustra.itch.io/imagegram

[2] https://dritchie.github.io/pdf/thesis.pdf

Show HN: MarkovJunior, PPL based on pattern matching and constraint propagation

ExUtumno — Wed, 01 Jun 2022 12:58:33 +0000

Article URL: https://github.com/mxgmn/MarkovJunior

Comments URL: https://news.ycombinator.com/item?id=31581158

Points: 36

# Comments: 14

New comment by ExUtumno in "Show HN: Wave function collapse algorithm"

ExUtumno — Sun, 02 Oct 2016 19:15:15 +0000

At first I thought that my methods don't offer anything new to text generation besides the Markov chain, but several people already proposed ideas that sound sensible, so let me know if you make anything!

New comment by ExUtumno in "Show HN: Wave function collapse algorithm"

ExUtumno — Sat, 01 Oct 2016 19:20:40 +0000

Most of the examples in the repo have those NxN all one color patches. Or, without (C2) the algorithm would have generated completely empty integrated circuits, or completely grass terrain, which is really boring.

You understood right, it's constraints + probabilities.

Btw, I have different algorithm that satisfies (C2) perfectly, but not (C1): https://github.com/mxgmn/ConvChain

New comment by ExUtumno in "Show HN: Wave function collapse algorithm"

ExUtumno — Sat, 01 Oct 2016 18:42:30 +0000

There are special approaches to generating music. The best for ratio of quality/complexity that I know of are Markov constraints https://www.youtube.com/watch?v=buXqNqBFd6E and WaveNet. I don't think WFC offers something useful and new for generation of music.

New comment by ExUtumno in "Show HN: Wave function collapse algorithm"

ExUtumno — Sat, 01 Oct 2016 15:48:32 +0000

In overlapping models we store probabilities for NxN blocks of colors/tiles. In non-overlapping models we store probabilities for individual colors/tiles.

New comment by ExUtumno in "Show HN: Wave function collapse algorithm"

ExUtumno — Sat, 01 Oct 2016 15:36:17 +0000

Yes, (C1) is a constraint problem. But we also want to satisfy (C2) as close as possible, otherwise we could have just colored some outputs in a single color.

New comment by ExUtumno in "Show HN: Wave function collapse algorithm"

ExUtumno — Sat, 01 Oct 2016 14:59:28 +0000

A very good question! The opposite of it is also important, can we follow some heuristics while creating tilesets to minimize contradiction rates, but not making tilesets easy? I don't know. If someone knows please tell me.

New comment by ExUtumno in "Show HN: Wave function collapse algorithm"

ExUtumno — Fri, 30 Sep 2016 21:05:31 +0000

We need to interpret those coefficients somehow. Real coefficients can be interpreted as mixing of colors, but for complex ones I don't see a good interpretation.

New comment by ExUtumno in "Show HN: Wave function collapse algorithm"

ExUtumno — Fri, 30 Sep 2016 20:57:35 +0000

I'm not experienced with the license law, but people told me that it's better to have license text in source files themselves, because I have samples in the repo that I have no idea who has rights for.

The license is MIT.

New comment by ExUtumno in "Show HN: Wave function collapse algorithm"

ExUtumno — Fri, 30 Sep 2016 20:53:37 +0000

I wonder too =). But it'll run like forever on a high res image. For high res image you want to use something like texture synthesis, see my reply to fitzwatermellow for more.

New comment by ExUtumno in "Show HN: Wave function collapse algorithm"

ExUtumno — Fri, 30 Sep 2016 20:46:58 +0000

If you use overlapping model (there are 2 models in the repo) with 1xN patterns, it would be a the same as (N-1)th order Markov chain.

New comment by ExUtumno in "Show HN: Wave function collapse algorithm"

ExUtumno — Fri, 30 Sep 2016 20:41:55 +0000

What do you mean by "code can be constructed with graphs"?

New comment by ExUtumno in "Show HN: Wave function collapse algorithm"

ExUtumno — Fri, 30 Sep 2016 20:35:34 +0000

So basically make a not-easy tileset with the shapes of Penrose tiles. Yes, this could be interesting.

New comment by ExUtumno in "Show HN: Wave function collapse algorithm"

ExUtumno — Fri, 30 Sep 2016 20:24:24 +0000

Thanks, I'll look into it.

New comment by ExUtumno in "Show HN: Wave function collapse algorithm"

ExUtumno — Fri, 30 Sep 2016 20:22:20 +0000

About harder and easier to satisfy, the question of how the rate at which the algorithm runs into contradictions depends on the input is not easy at all. There is no simple correlations between the contradiction rate and the size of the input.

But the first thing you'll notice if you feed it an image with a lot of patterns, is that it will work very slowly.

Yeah, the corpus thing can be done if we cut out rare patterns and leave only frequent ones. I haven't tried it though.

New comment by ExUtumno in "Show HN: Wave function collapse algorithm"

ExUtumno — Fri, 30 Sep 2016 17:44:55 +0000

Thanks!

Yeah, you a right, I'll upload slower gifs. Right now youtube video has the slowest speed, in fact it has segments with no frame-skipping at all: https://youtu.be/DOQTr2Xmlz0