Tag
The developer behind the live evolution sim EDEN evolved its creatures' tiny neural network brains offline for 300 generations, raising win rates on previously unseen rounds (e.g., Ravagers from 7% to 62%, Nomads from 42% to 95%). The key fix was that stream founders previously got random brains each round, discarding evolved wiring; co-evolution across species is the next step.
This preprint introduces Neuroevolution Arena, a GPU-accelerated artificial life platform for comparing evolution and RL-based update regimes across neural architectures via a nested ecological evaluation protocol.
This paper proposes a neuroevolution-based fine-tuning method to improve the accuracy of quantized deep learning models, showing that nearest-neighbor rounding alone is suboptimal and that evolutionary mutation of weights can yield better results on architectures like VGG and ResNet.