-
Book Overview & Buying
-
Table Of Contents
Architecting AI Software Systems
By :
We have made the case for how to build AI-enabled systems. Building AI-enabled systems is a challenge. There exist high expectations for the system and the likelihood that the software system is complex. There are several ways for a software development effort to fail; these range from not understanding key requirements, a technology not performing as expected, an erroneous system design, and compute, storage, or data flows being misunderstood. A major driver for failure is that humans can lose trust in the system if outputs are inconsistent, wrong, or just not sensible. AI-enabled systems have these risks plus algorithmic complexity, sensitive to off-nominal scenarios, and data inputs being off. A key driver of AI system failures is that users lose trust or have low confidence in the results provided by the system. AI-enabled systems must be built from inception so that they will utilize AI technologies. These systems must address decision-making without...
Change the font size
Change margin width
Change background colour