Skip to content

Latest commit

 

History

History

Folders and files

NameName
Last commit message
Last commit date

parent directory

..
 
 
 
 

Scenario 5 Task Decomposition

Title

Machine Intelligence with Knowledge Graphs.

Description

A hybrid system generates diagnoses (Scenario) based on the patient’s symptoms and medical history stored in a Knowledge Graph. A human operator and an embeddings-based (ProcessingMethod) system (both Actors) with complementary scientific expertise (Capability) perform the prediction and justify their propositions to each other (InteractionTask) in order to come up with a final decision (EndGoal)

Visual Task Decomposition

S5.png

Terminology

(dynamic) symptoms, embeddings, historical medical events, diagnoses, justifications, decisions. (static) human memory, medical KG

Inference Steps:

  1. (human) recollect similar events ;
  2. (human) induce the diagnosis;
  3. (machine) train and
  4. (machine) classify diagnoses;
  5. generate a justification for the diagnosis and
  6. idem
  7. present it to the other agent that can
  8. idem
  9. obtain the information;
  10. ultimately, generate a combined decision.

Tasks

Prediction (1-4), Explaining (5-6), Team Role Allocation (7-10)