Epilepsy seizure detection and prediction using techniques such as deep learning methods

An apparatus, comprising:

• one or more memories comprising program instructions;
• one or more processors, the one or more processors causing, in response to retrieval and execution of the program instructions, operations comprising:
• performing one or both of epilepsy seizure detection and prediction at least by performing the following:
• running a plurality of input signals from sensors for epilepsy seizure detection through a plurality of classification models;
• and
• applying weights to outputs of each of the classification models to create a final classification output;
• adjusting the weights to tune relative output contribution from each classifier model in order that accuracy of the final classification output is improved, while power consumption of all the classification models is reduced;
• and
• performing the one or both of epilepsy seizure detection and prediction with the adjusted weights.

A method, comprising:

• accessing input signal streams from multiple sensors for epilepsy seizure detection;
• performing processing on the input signal streams to convert to processed data having one or more representations suitable for classification tasks;
• streaming processed input signal streams into a plurality of classification models for one or both of epilepsy seizure detection and prediction and performing training to create multiple classification models having trained network weights that are not adjustable after training;
• and
• sending information defining the plurality of classification models with their trained network weights that are not adjustable to one or more wearable computer platforms for use by the one or more wearable computer platforms to perform one or both of epilepsy seizure detection and prediction.