Synthesizing Disparate Experiences in Episodic Planning
Abstract
Many decisions are actually made by synthesizing previous experience. Often, this involves many different experiences coming together to form a feasible solution. This paper presents a statistical model for predicting the outcome of solutions based on multiple experiences. In edge organizations, such as emergency first responders, it often requires the expertise of more than one person to form an approach to a complex problem. Unfortunately, each planner only has access to his or her own memories. We propose to use an artificial intelligence decision aide to help bridge this gap, by reasoning over distributed collections of previous experiences. The key research questions that we address include: How can an artificial reasoner form a plan based on several disparate experiences from different sources? How can we gauge the potential efficacy of such a plan? How can we trust this plan if a clear line cannot be drawn to one author? We will also discuss such critical issues as analogies in planning with disparate experiences, civil-military planning by analogy, trust, provenance, and organizational issues in planning.
Document Details
- Document Type
- Technical Report
- Publication Date
- Jun 01, 2008
- Accession Number
- ADA486977
Entities
People
- Anthony J. Ford
- James H. Lawton
Organizations
- Rome Laboratory