Online Learning and Continual Training

A model trained on last week's data may already be stale. User preferences change, new products appear, fraud patterns shift, and seasonal effects move faster than a weekly batch job.

Online learning and continual training address this by updating the model from fresh data instead of waiting for the next full retrain. The hard part is doing that without turning the model into an unstable feedback loop.