Improving results
Handling improvements
No model will be perfect. Once we release a model for others to use, we need to think about how we are able to support it. Other users could use it for retraining, adapting it to their data, if it doesn’t work straight away. There are also pipelines where systems can build problematic images back into the training set to help improve the performance next time. Perhaps future collaborations could allow you access to more annotated training data from other users. In this video, we discuss options for improving model performance once released.
Experimental Design for Machine Learning
Reach your personal and professional goals
Unlock access to hundreds of expert online courses and degrees from top universities and educators to gain accredited qualifications and professional CV-building certificates.
Join over 18 million learners to launch, switch or build upon your career, all at your own pace, across a wide range of topic areas.
