-
Book Overview & Buying
-
Table Of Contents
Architecting AI Software Systems
By :
As has been stated, the building of an AI system has as its core the building of components that are expected to make decisions. The decisions that are to be made rely heavily or almost entirely on the data that comes into the system. A decision is only as good and valid as the data that was used. The next sections describe key tasks that can be used to ensure the highest quality of data enters the system. These tasks can be both tedious and challenging since, for the initial system development, a human is needed. That said, for further system development, these tasks can be done in an automated manner with checking and alerting.
Understanding data quality is critical for effectively training, tuning, and maintaining AI pipelines. Quality checks should be rigorously configuration-controlled and tested, with minimum requirements explicitly specified. These include comprehensive assessments of data completeness to ensure...
Change the font size
Change margin width
Change background colour