TAIBOM Architecture

Fundamentally, TAIBOM is a method of describing a complex information systems (AI), with complex dependencies. It provides a method of annotating this system with "risk modifiers ". This allows a TAIBOM user to dynamically reason about the trustworthiness of the system

More technically, its a distributed knowledge graph, where each entity and each relationship can be authored/contributed by a different actor. Each relationship and each actor is encapsulated with a verifiable credentials

The core of the architecture is the definition of the verifiable credential schemas. These are formally specified here http://localhost:3000/schemas/ .

This base set defines the minimum needed to usefully describe risk relations in complex AI systems. They are inherently extensible.

The second critical aspect is the information flows. What physical process do we use to aggregate the descriptors. These are explained in the infomatoin flow section

Finally we outline how you reason over the descriptors.