AI
Implemented — in the honest sense: AI demonstration examples
written in ordinary Karkain arithmetic exist under examples/09-ai/ and run
byte-identically on both engines.
There is no ``std.ai`` module and no AI/ML framework surface in the language. These examples show classic ML geometries (distance, weighted votes, linear decision boundaries) implemented by hand with plain arrays, maps and arithmetic — deliberately, so the language itself can be exercised without pretending a framework exists. See Implemented for how the status vocabulary is used.
File |
Demonstrates |
|---|---|
|
Manhattan-distance nearest-neighbour classification by hand |
|
Perceptron-style weight updates and a decision boundary |
K-nearest neighbour (excerpt)
func distance(a, b) {
return abs(a[0] - b[0]) + abs(a[1] - b[1])
}
...
Run:
$ karkain run examples/09-ai/01_nearest_neighbor.kark
See also
Machine Learning — the adjacent hand-rolled ML category. Planned — the not-yet-implemented AI framework roadmap.