| CPC G06N 5/022 (2013.01) [G06F 16/258 (2019.01)] | 20 Claims |

|
1. A method for managing training data, comprising:
storing, in response to a determination that new training data is collected by a sensor, the new training data into a collected data stream of a storage pool in a processor-based machine learning system;
storing, in response to a determination that the new training data and historical data stored in a full data stream of the storage pool are refined into refined training data utilizing a dataset distillation algorithm, the refined training data into a refined data stream of the storage pool;
storing the new training data into the full data stream of the storage pool; and
providing, in response to receiving a use request for training data from at least one of an edge device and a cloud, training data from at least portions of one or more of the collected data stream, the full data stream and the refined data stream.
|