Skip to content

aintoniodev/genetic-programming-pipelines

v1.0.0MIT

Genetic programming (GP / programación genética) applied to neural-network and deep-learning pipelines: evolutionary program-space search (code, trees, configs, architectures, augmentations, prompts, agents) with fitness-ladder design, LLM operators, champion selection, diversity pressure, and objective audits. Activates only on genetic-programming topics, not generic autoresearch.

GitHub
genetic-programming-pipelines

Genetic programming (GP) / programación genética applied to neural-network and deep-learning pipelines: evolutionary search in program space (code, trees, configs, architectures, augmentations, prompts, agents) and fitness/evaluation design for GP loops. Covers search-space framing (what evolves vs what gradients handle), fitness ladders with rank-inversion validation, LLM-driven genetic operators with deterministic gates, champion selection, diversity pressure, objective audits, lineage logging. Activates ONLY when the topic is genetic programming — ES: 'programación genética', EN: 'genetic programming'/'GP' — NOT for generic autoresearch, HPO, or NAS without a genetic-programming framing.