Learn how to evaluate an AutoML Vision model using AUC and other built-in model quality metrics.
- [Instructor] A critical aspect of working…with AutoML for vision is model evaluation.…The model has to produce the correct labels.…There are several ways that model quality is measured.…You remember that the data is split…into training, test, and holdout for quality evaluation.…And these metrics are presented in a nice interface.…So what are these?…There's basically four types of metrics.…The first is the AUC, or area under the precision…or recall curve, and sometimes called AuPRC.…
And to understand that, you have…to understand precision and recall.…Precision is the ability for the model…to predict fewer false positives.…Recall is the ability for the model…to predict fewer false negatives.…You use what's called a confusion matrix…which will show you the percentage of incorrect…predictions, either false positives or false negatives.…Now this only is available as of this…recording for single-label models.…
So what does this look like?…Here is an example of two different…confusion matrices for our architecture labeling.…
- Hosting options: Serverless, containers, and virtual machines
- Enabling the GCP ML AIs
- Preparing data with Cloud Dataflow and Dataprep
- Modeling predictions for images, video, text to speech, and cloud translation
- Machine learning with AutoML
- Advanced machine learning and deep learning
- Machine learning architectures
Skill Level Intermediate
1. Machine Learning on Google Cloud Platform
2. Machine Learning API Services
3. Machine Learning with AutoML
Understand AutoML Vision4m 57s
4. Advanced Machine Learning
5. Machine Learning Architectures
Next steps1m 30s
- Mark as unwatched
- Mark all as unwatched
Are you sure you want to mark all the videos in this course as unwatched?
This will not affect your course history, your reports, or your certificates of completion for this course.Cancel
Take notes with your new membership!
Type in the entry box, then click Enter to save your note.
1:30Press on any video thumbnail to jump immediately to the timecode shown.
Notes are saved with you account but can also be exported as plain text, MS Word, PDF, Google Doc, or Evernote.