*********************************
There is now a CONTENT FREEZE for Mercury while we switch to a new platform. It began on Friday, March 10 at 6pm and will end on Wednesday, March 15 at noon. No new content can be created during this time, but all material in the system as of the beginning of the freeze will be migrated to the new platform, including users and groups. Functionally the new site is identical to the old one. webteam@gatech.edu
*********************************
Title: Spiking Neural Network for Unsupervised Learning of Spatiotemporal Patterns
Committee:
Dr. Mukhopadhyay, Advisor
Dr. Raychowdhury, Chair
Dr. Rozell
Abstract: The objective of the proposed research is to achieve learning of complex spatiotemporal patterns with spiking neural network. We present a heterogeneous spiking neural network (H-SNN) as a novel, feedforward SNN structure capable of learning complex spatiotemporal patterns with spike-timing-dependent plasticity (STDP) based unsupervised training. We also present a GPU accelerated platform for convolutional SNN with dynamic network structure, which facilities the development of H-SNN. Within H-SNN, hierarchical spatial and temporal patterns are constructed with convolution connections and memory pathways containing spiking neurons with different dynamics. We demonstrate the formation of long and short term memory in H-SNN and distinct response functions of memory pathways. In simulation, the network is tested on visual input of moving objects to simultaneously predict for object class and motion dynamics. Results show that H-SNN achieves prediction accuracy on similar or higher level than supervised deep neural networks (DNN). Compared to SNN trained with back-propagation, H-SNN effectively utilizes STDP to learn spatiotemporal patterns that have better generalizability to unknown motion and/or object classes encountered during inference. In addition, the improved performance is achieved with 6x fewer parameters than complex DNNs, showing H-SNN as an efficient approach for applications with constrained computation resources. We then demonstrate an event-driven SNN for efficient processing of data from neuromorphic vision sensors, which achieves processing speed improvement and reduces distortion to low-precision networks, leading to better learning performance than discrete-time simulated network.
