Oscillator associative memories for high-capacity, compositional inference

We introduce a novel, high-capacity associative memory capable of factorizing compositional representations of the world. The approach is implemented as a continuous-time oscillator neural network. By performing factorization with a continuous-time dynamical system, the network provides efficient solutions to computationally hard problems such as perceptual inference world variables such as form and motion, or combinatorial optimization problems such as subset sum. We demonstrate favorable performance compared to existing approaches to factorization, efficient implementation, and relevance to standard tasks such as the subset sum problem. We also identify concrete reasons why our model exhibits improved capacity, by formulating methods to track its convergence.