From the course: Agentic AI Design Patterns for GenAI and Predictive AI
Join today to access over 25,500 courses taught by industry experts.
Proactive retraining
From the course: Agentic AI Design Patterns for GenAI and Predictive AI
Proactive retraining
- One of the biggest concerns and challenges with any AI system is when its output becomes incorrect, because the real world data it's seeing has changed either through data drift, where the new data is now different from the data used to train the model or concept drift, where the relationship between the data in the real world has changed compared to how the model was trained. We've covered data and concept drift in the previous foundational courses, so I won't further describe them in this video. The fact is that they are a very real problem for many AI systems. But one of the reasons that it's such a problem is because it can take us a while to notice that drift is occurring, which means that for a time, the output data that is being received and used will be flawed. Here we show a simple scenario where an AI engineer eventually realizes that a predictive AI system has been producing flawed or substandard…
Download courses and learn on the go
Watch courses on your mobile device without an internet connection. Download courses using your iOS or Android LinkedIn Learning app.