Skip to content

Releases: pkalivas/radiate

v1.2.6

15 Dec 22:13
Compare
Choose a tag to compare

This release is primarily filling in gaps with documentation and some code cleanup.

Also included are impl Integer<T> for u8, u16, u32, u64, and u128 so these types can be used with the IntGene.

Full Changelog: v1.2.5...v1.2.6

v1.2.5 Expression Trees (Genetic Programming), Multi-Objective Optimization

01 Dec 00:15
Compare
Choose a tag to compare

Adding support for genetic programming through tree structures, commonly known as Genetic Programming or Expression Trees.
Adding support for multi-objective optimization and Pareto fronts.

New selectors for multi-objective optimizations: NSGASelector - Non dominated sorting.

Refactoring to use specific chromosomes for each gene in order to to support the NodeChromosome for Graphs and Trees.

Adding doc page for better understanding of the library: https://pkalivas.github.io/radiate/

RC 1.2.2 Base re-write.

18 Nov 23:55
Compare
Choose a tag to compare

Finalizing the base re-write. Publishing to crates. Also included is radiate-extensions, extensions to the core library which include genetic programming and a graph data structure built for evolving. This graph can represent neural networks with implementations of directed weighted graph (dense) layers, cyclical weighted graphs (recurrent), gated recurrent units, long short term memory units, attention units. Any type of graph can be built. Also included in the extensions is a base impl for regression analysis used for graph fitness scoring.

updates to README.md and documentations with more examples and library walk throughs to come.

This will serve as the base for future updates.