@enasmel and myself are excited to announce our paper "Meta-Learning through Hebbian Plasticity in Random Networks" arxiv.org/abs/2007.02686
Instead of optimizing the neural network's weights directly, we only search for synapse-specific Hebbian learning rules. Thread 👇
Instead of optimizing the neural network's weights directly, we only search for synapse-specific Hebbian learning rules. Thread 👇
Starting from completely random weights, the discovered Hebbian rules enable an agent to navigate a dynamical 2D-pixel environment; likewise they allow a simulated 3D quadrupedal robot to learn how to walk in around 40 timesteps in the absence of any explicit reward.
The random Hebbian network is also able to adapt to damages in the morphology of the quadrupedal robot, while a fixed-weight network fails to do so. 
Interestingly, it's even possible to zero out many of the weights and the network is able to recover in a completely self-organized way. This experiment was inspired by @zzznah's work on self-organizing cellular automata and earlier work on random Hebbian networks by @DFloreano
We hope this work might also be of interest to others working on biologically inspired neural networks @neurograce @TonyZador @tyrell_turing @kenneth0stanley @jeffclune @asoltoggio @ThomasMiconi @KordingLab @NeuroAILab @hardmaru @Jack_W_Lindsey.
Code will be published soon.
Code will be published soon.
