A Spiking Neural Learning Classifier System
This is an incremental improvement for reinforcement learning applications, particularly in simulated robotics.
The paper tackles the problem of traditional Learning Classifier Systems (LCS) failing to solve certain continuous-input problems efficiently by introducing a spiking neural network representation with constructivist growth and temporal state decomposition, resulting in optimal performance where traditional methods struggle.
Learning Classifier Systems (LCS) are population-based reinforcement learners used in a wide variety of applications. This paper presents a LCS where each traditional rule is represented by a spiking neural network, a type of network with dynamic internal state. We employ a constructivist model of growth of both neurons and dendrites that realise flexible learning by evolving structures of sufficient complexity to solve a well-known problem involving continuous, real-valued inputs. Additionally, we extend the system to enable temporal state decomposition. By allowing our LCS to chain together sequences of heterogeneous actions into macro-actions, it is shown to perform optimally in a problem where traditional methods can fail to find a solution in a reasonable amount of time. Our final system is tested on a simulated robotics platform.