Offline Learning In Ai
A machine learning approach where the model is trained using a finite, static dataset. The model is trained over the complete dataset in one or several passes, and the parameters are updated after the learning process has been completed.
Areas of application
- Natural language processing
- Computer vision
- Recommendation systems
- Predictive maintenance
- Fraud detection
Example
Training a recommendation system on a catalog of products sold by an online retailer without any real-time data or feedback from customers.