Excerpt from Enterprise Architecture Fundamentals:
That learning loop can be used to leverage the benefits of AI and ML technologies across enterprise architecture:
Observation is the main locus of DL technologies, with advances obtained through a shift from Supervised to RL. Beyond its economic benefits (from down- sizing the armies of tutors who redundantly put names on pictures), the significance of RL comes from detaching learning from human common sense. But cutting off common sense from Machine learning relies on the implicit assumption that meanings can be extracted from raw data like gold nuggets from river beds, independently of what humans may think. That (literally) open-minded approach may be a boon for the discovery of emerging consumers’ fancies, but not so for the development of well-thought-out strategies. More generally, the difference is between pattern matching, open to discovery, and policy making, focused on purpose: when policies are considered, Deep learning must be combined with the reasoning capabilities of Knowledge graphs.
Reasoning is the realm of KG technologies, whose ubiquity can be explained by one technical rationale and two functional ones (figure 16-3):
- The implementation of KG with neural networks enables a seamless integration with DL applications.
- Serving as semantic networks, KG provide modular and versatile interfaces with natural languages.
- Serving as property graphs, KG provide robust interfaces with systems modeling languages (in particular, relational models). They can then combine with DL to mine data from applications and databases.
For EA, the main benefit of KG is their seamless integration with ontologies and, consequently, their ability to leverage learning across the full range of systems, organization, and business representations.
Judgment represents the ultimate yardstick of learning; at the enterprise level, that yardstick is both collective and individual:
- Collective, because judgments should be backed by transparent and traceable knowledge and reasoning, independently of individual beliefs or opinions
- Individual, because when judgments translate into decisions, even ones taken collectively, their traceability also implies personal accountability
That pivot from learning (individual and collective knowledge) to decision- making (individual accountability) is arguably a critical issue for organizations. It can be analyzed with regard to the problem at hand, or from the broader perspective of organizational behavior.