Genetic Programming
Also known as GP
Overview
Genetic programming searches for programs by repeatedly selecting, mutating, and recombining candidate trees or graphs according to their performance. A major use is symbolic regression, where it seeks interpretable equations; other applications include scheduling and routing heuristics, feature construction, circuit design, and control policies. Current approaches combine multi-objective search and surrogate evaluation with GPU acceleration or language-model guidance to explore candidates more efficiently. A central challenge is balancing search cost and solution quality with generalization and simplicity, since highly fit programs can be complex or fail on new data.
Latest papers
61 papers3 papers in the last four weeks, down 40% on the four weeks before. 0.0% of all new papers.