DATA SCIENCE APPROACHES TO AUTOMATION OF ANALYTIC WORK FLOWS
Abstract
Data Science Approaches to Automation of Analytic Work Flows The objective of the research program described in this proposal is to define a system engineering approach to the functional design of a closed-loop adaptive information collection system supportive of both intelligence analysis and tactical mission operations. We propose to achieve this objective by accomplishing the following related goals: [Analytics]: careful assessments of ontological elements and data technologies supporting current-day IC/tactical analytic frameworks that yield recommendations for advancements in analytic methods in both domains, and [Architecture]: research on the ontological elements and data technologies supporting an optimal cognitively-based software framework for a closed-loop adaptive information collection system; the focus is on design of a computational framework of support to both intelligence analysis and mission operations. The proposed research program is largely motivated by critical shortfalls in various quality criteria in today’s information environment for IC/tactical applications, as enumerated in [Rohr, Miller, 2016]: Ambiguous vocabularies, inaccuracies, interoperability shortfalls, lack of accurate methods for information linking, of provenance tracking throughout the system architecture, and of governance of the overall architectural framework. Our goal is to provide a top-level functional design, for a collection system that supports iterative knowledge evolution helpful to the IC and tactical mission domains, following the methodology documented in [Arp, 2015] and drawing on standard public domain resources developed by the World Wide Web Consortium (W3C) including the Web Ontology Language (OWL) and associated software. This methodology, in brief, consists in translating assertions of different sorts – including free text assertions – into the computable format provided by OWL in such a way that the content of these assertions becomes available for automated query and analysis. The proposed effort is split into two parts, in such a way that interdependencies among these parts are exploited to yield a single overarching design: 1) a specification of both the representative data involved and of a functional ontological framework that provides the semantic clarity required along with the agility to adapt to new types and classes of data; 2) a systems engineering design activity that integrates the ontological framework, collection operations, and the set of analytic methods into a holistic functional system design. The proposed program will have the Principal Investigator as Dr. James Llinas, an Emeritus Research Professor at the University at Buffalo, with extensive credentials in the national intelligence systems design and data and information fusion domains central to this effort. Other key personnel will include Drs. Barry Smith and Alex Nikolaev of the University at Buffalo, and Mr Ron Rudnicki of CUBRC who has extensive real-world experience in applying semantics and ontological science to real systems; focused student support will also be provided. The project is expected to have considerable public benefit in specifying robust systemic approaches to high- quality adaptive information collection, a widely-demanded capability for many civil and business applications, as well as in the transference of advances in semantic science to the public sector.
Document Details
- Document Type
- DoD Grant Award
- Publication Date
- Mar 22, 2018
- Source ID
- N002441810003
Entities
People
- James Llinas
Organizations
- Research Foundation for the State University of New York