Comments (5)
Some observations about the "Separate by shapes" tasks:
- The matrices tend to be big in all cases except two (777 most importantly), all the matrices are at least 9x9.
- It is worth to do some preprocessing to determine wheter we want to consider multicolor Shapes or just normal Shapes (obviously, only non-background ones). Check whether there is the same amount of (multicolor) Shapes in the input and the output. Also with diagonals.
- How shall we decide how to modify the output? Shapes that are in the same position or are of the same color seem reasonable approaches, but not sure if this will be enough. How to choose the output in tasks like 429, 602 or 750, for example?
- If the number of Shapes is the same in the input and in the output, then consider positions to select the output Shapes. This will not catch tasks like 396 or 517, but I think it's the best I can do.
- The number of Shapes in a matrix is always low. Haven't seen any case with more than 7.
from arc_kaggle.
I uploaded a first version in this commit: 583dbce.
For now, it deals with most of the "separate by shapes" cases. It solves tasks like 496, 528, 611 or 681, amongst several others.
I still need to improve the way of choosing the final best three candidates and deal with the "separate by colors" tasks.
from arc_kaggle.
Next improvement should be: if there is some fixedColor, then only consider it if it is the only color (apart from the background color)
from arc_kaggle.
For separating by color, I should:
- Consider colors that are not fixed and, are not background and do not vanish. In general, not consider common colors
- Consider tasks in which each sample has at least one of such colors, and at least one sample contains two of such colors
from arc_kaggle.
Separation by colors done in 2934eaa. Closing the issue.
from arc_kaggle.
Related Issues (20)
- can detectFrontier function detect a single row as a frontier?
- the huge gap between train/eval and test HOT 1
- Add function ifTestInTrain HOT 2
- Implement EvolvingLine HOT 6
- "Fill the blank" tasks with several blanks HOT 1
- Check out task 625 HOT 3
- Avoid repeating tryOperations if the task looks the same HOT 1
- Do we need to re-think the initial recoloring? HOT 1
- Task 785 is not solved HOT 3
- Tasks with cropAllBackground+doSomething+recoverBackground HOT 2
- Ignore asymmetric grid if it makes sense HOT 1
- Deal with task 645 HOT 2
- cropManyShapes function? HOT 6
- Connect pixels with diagonals HOT 1
- Color symmetric pixels HOT 1
- Connect/extend in grid cells
- Final quick wins HOT 7
- Deal with important colors that are not considered important
- Correct fixedColors AND commonChangedInColors HOT 1
Recommend Projects
-
React
A declarative, efficient, and flexible JavaScript library for building user interfaces.
-
Vue.js
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
-
Typescript
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
-
TensorFlow
An Open Source Machine Learning Framework for Everyone
-
Django
The Web framework for perfectionists with deadlines.
-
Laravel
A PHP framework for web artisans
-
D3
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
-
Recommend Topics
-
javascript
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
-
web
Some thing interesting about web. New door for the world.
-
server
A server is a program made to process requests and deliver data to clients.
-
Machine learning
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
-
Visualization
Some thing interesting about visualization, use data art
-
Game
Some thing interesting about game, make everyone happy.
Recommend Org
-
Facebook
We are working to build community through open source technology. NB: members must have two-factor auth.
-
Microsoft
Open source projects and samples from Microsoft.
-
Google
Google ❤️ Open Source for everyone.
-
Alibaba
Alibaba Open Source for everyone
-
D3
Data-Driven Documents codes.
-
Tencent
China tencent open source team.
from arc_kaggle.