-
Book Overview & Buying
-
Table Of Contents
Architecting AI Software Systems
By :
The heart of an AI system is the learning component – the place where the “smarts” of the system reside. Matching the right technique to customer objectives drives the technical teams’ implementations. Many times, an architecture must be able to support different types of techniques to solve a specific problem: the importance of matching the relevant technique to a customer domain and the importance of visualization, and how the architecture should be used for capturing model performance metrics.
It is important to address model quality monitoring to aid in troubleshooting and monitoring of the system. The use of canaries and robust gold standard testing is key so that the decisions that are driven by the model are correct.
The authors are quite confident that many more techniques will be created and deployed. It is the charge of the architect to design and build robust systems to deal with this uncertainty. The aspiring...
Change the font size
Change margin width
Change background colour