Self-assembling Neural Networks Can Open New Directions For Ai Research

Scientists at the IT University of Copenhagen have proposed the concept of self-assembling neural networks that “grow through a developmental process that mirrors key properties of embryonic development in biological organisms”. This paper outlines a vision to develop an AI system where neurons self-assemble, grow, and adapt based on the task at hand, emulating the natural processes observed in biological networks. The proposed graph neural network (GNN) encoding involves two interconnected neural networks, a policy network and a generative network.