PROGRAMS BY home.arcor.de/w.lorenz65/mlbench

A suite of machine learning benchmarks where each agent must solve a lot of different tasks without recompilation. This means that the programmers cannot manually specify topologies or adjust parameters to specific tasks.