Skip to content

conceptual modeling for kgrid

ajflynn edited this page Sep 18, 2016 · 11 revisions

Why do we need a conceptual model?

The Knowledge Grid (KGRID) is a complex system with many components and potential components. We have used and will continue to use a variety of metaphors to describe the KGRID. Perhaps the most consistent metaphor we have applied so far is the metaphor of a physical Library. However, we have also relied on infrastructure metaphors (bridges and highways, the telephone system, the electrical power grid, cold storage, water reservoirs and distribution pipes, etc.) An evolving conceptual model that can help us simultaneously and continuously confirm/reconfirm and challenge/refute the validity of these metaphors as they apply to the KGRID could help us communicate as a team and share our work with others.

The main concepts we are dealing with all the time are these . . .

Learning Health System

A Learning Health System is a health system that (i) makes all consenting consumers' health data avaialble to study and learn from, (ii) makes best practice knowledge immediately available to support decisions, (iii) demonstrates continuous improvement through ongoing learning, (iv) does i, ii, and iii in ways that are both routine and economical using infrastructures suited to its purpose, and (v) nurtures and develops a culture of learning to sustain and grow itself.

Learning Cycle

A Learning Cycle is a framework for an empirical method of continuous, routine learning. The Learning Cycle framework focuses attention on a Problem of Interest. A person or persons Decides to Study a Problem of Interest to them. The framework lays out a series of steps once a Decision to Study has been taken. First, data are collected. Next, data are aggregated. Data are then Analyzed. Interpretation of results follows. The reason to interpret results is to see whether anything has been learned. If something has been learned then knowledge representation is the next step. Once knowledge is represented it is then stored somewhere. Next, knowledge on hand is used to generate individualized, tailored messages to inform specific decisions being faced by one or more people. Once messages are generated they are communicated. Messages are then received and acted on. Finally, the actions taken on the messages are discerned and become the source of new data. The cycle can then be repeated again by collecting those new data.

K-Component

The K-Component, named after the K in Knowledge, is an architectural component to support people who do the work of completing Learning Cycles. Specifically, the K-Component provides Knowledge Object management and discovery capabilities on top of a storage mechanism for machine-interpretable, encoded knowledge.

P-Component

The P-Component, named after the P in Practice, is an architectural component to support people who do the work of completing Learning Cycles. Specifically, the P-Component provides Knowledge Object access, registration, and notification on top of a storage mechanism for machine-interpretable, encoded knowledge.

Grid

A Grid is an interconnected set of 2 or more libraries. We envision a Grid of libraries that connects organizations with a desire to share knowledge openly for the purpose of improving human health.

Library

A Library is an actual instance of our K-component architecture running in software at a specifiable location.

HomeShelf

A HomeShelf is an actual instance of our yet-to-be-developed HomeShelf architecture running in software at a specifiable location.

Collection

A Collection is a group of objects stored or held in a Library or in a Shelf.

Knowledge Object

A Knowledge Object is a digital object with a binary payload of some executable, machine-interpretable software code combined with a variety of descriptive information, most of which is serialized in RDF.

Payload

A Payload is a binary file containing some executable, machine-interpretable software code.

ObjectTeller

ObjectTeller is an application targeted at supporting users who wish to manage a Library.

Client application

A Client Application is any application that uses actual K-Component, P-Component, or HomeShelf architectures running in software to do something. Therefore, ObjectTeller is a Client Application. Sometimes we use this term to refer to applications that are not ObjectTeller however that usage does not reflect its precise meaning.

External application

An External Application is a class of Client Applications that have NO capabilities to control Library or HomeShelf management functions. External Applications can only interact with a Library of HomeShelf through exposed APIs that enable capabilities unrelated to managing a Library of HomeShelf. External Applications may consume and interact with Knowledge Objects held in a Library or a HomeShelf.

Ideas (assertions-to-be-tested) and confusions (open-ended questions) about a conceptual model involving the concepts above

Ideas

  1. More components need to be specified to bring about a supporting infrastructure spanning from knowledge representation to message communication within the Learning Cycle framework

Confusions

1: If the P-Component is for Knowledge Object access, registration, and notification, should it be called the A-Component for Access?

Clone this wiki locally