System and method for generating data sets for model retraining

Abstract

A computer system is provided and programmed to assemble a plurality of synthetic datasets and mix those synthetic datasets into a synthetic dataset. An evaluation is then performed to determine whether an existing model should be associated with the synthetic data set or a new model should be trained from the existing model using the synthetic data set.

Description

[0001] Cross reference(s) to related application(s) [0002] This application claims priority to U.S. Application No. 62 / 847,621, filed May 14, 2019, and U.S. Application No. 15 / 931,369, filed May 13, 2020, the entire contents of which are hereby Incorporated by reference. technical field [0003] The techniques described herein involve processing and / or preparing datasets for retraining or training a model. More particularly, the techniques described herein relate to transfer learning techniques for machine learning, and more particularly deep learning. Background technique [0004] Machine learning and deep learning can give computers the ability to "learn" a specific task without explicitly programming the computer for that task. This can be achieved by developing a model (eg, a mathematical model) based on initial data and then using the model to evaluate new data. For example, a model could be developed to automatically recognize and distinguish between pictures cont...

AI Technical Summary

Problems solved by technology

However, generating (e.g., training) an entirely new model can be a resource-intensive process

Images