Anomaly detection with spiking neural networks
| DWPI Title: Computer-based method for detecting anomalies in dataset, involves outputting original value of input received by input layer by adaptive median filter (AMF) layer when none of median absolute difference (MAD) values exceeds threshold |
| Abstract: Detecting anomalies with a spiking neural network is provided. An input layer receives a number of inputs and converts them into phase-coded spikes, wherein each input is contained within a number of progressively larger neighborhoods of surrounding inputs. From the phase-coded spikes, a median value of each input is computed for each size neighborhood. An absolute difference of each input from its median value is computed for each size neighborhood. A median absolute difference (MAD) of each input is computed for each size neighborhood. For each input, an adaptive median filter (AMF) determines if a MAD for any size neighborhood exceeds a respective threshold. If one or more neighborhoods exceeds its threshold, the AMF outputs the median value of the input for the smallest neighborhood. If none of the neighborhoods exceeds the threshold, the AMF outputs the original value of the input. |
| Use: Computer-based method for detecting anomalies in dataset using SNN (claimed). |
| Advantage: The intermittent occurrence of spikes gives SNNs with lower energy consumption than other types of neural networks. The neuron is prevented from indefinitely retaining energy, which does not match the behavior of biological neuron. |
| Novelty: The method involves receiving a number of inputs (210) by an input layer in a spiking neural network. Each input is contained within a number of progressively larger neighborhoods of surrounding inputs. The inputs are converted into phase-coded spikes by the input layer. A median value of each input for each size neighborhood is computed from the phase-coded spikes by a first median layer. An absolute difference of each input is computed from the median value for each size neighborhood by an absolute difference layer. A MAD value of each input for each size neighborhood is computed from absolute differences, by a second median layer. A determination is made for each input by an AMF layer, when a MAD value for any size neighborhood exceeds a respective threshold. The median value of the input for the smallest size neighborhood is output by the AMF layer, when a MAD value of neighborhoods exceeds the threshold. |
| Filed: 6/10/2019 |
| Application Number: US16436744A |
| Tech ID: SD 14754.0 |
| This invention was made with Government support under Contract No. DE-NA0003525 awarded by the United States Department of Energy/National Nuclear Security Administration. The Government has certain rights in the invention. |
| Data from Derwent World Patents Index, provided by Clarivate All rights reserved. Republication or redistribution of Clarivate content, including by framing or similar means, is prohibited without the prior written consent of Clarivate. Clarivate and its logo, as well as all other trademarks used herein are trademarks of their respective owners and used under license. |