SRI Work on Science Taxonomies Jeffrey Alexander, Ph.D. Senior Science & Technology Policy Analyst Center for Science, Technology & Economic Development August 2011
Who We Are SRI is a world‐leading independent R&D organiza>on • Founded by Stanford University in 1946 – A nonprofit corpora>on – Independent in 1970; changed name from Stanford Research Ins>tute to SRI Interna>onal in 1977 • Sarnoff Corpora>on acquired in 1987 (formerly RCA Laboratories) will become fully integrated into SRI effec>ve January 1, 2011 SRI headquarters, Menlo Park, CA • 2,000 staff members combined – 800 with advanced degrees – More than 20 loca>ons worldwide • Consolidated 2010 revenues: $500+ million Sarnoff headquarters, Princeton, NJ SRI St. Petersburg, Florida SRI Harrisonburg, Virginia SRI State College, Pennsylvania SRI Tokyo, Japan SRI Washington, D.C. 2
Center for Science, Technology & Economic Development Sample Ac>vi>es • Science, technology, and innova>on measurement and indicators (NSF, Saudi Arabia) • State/regional economic and innova>on strategies (Florida, Virginia) • Research ins>tute planning and evalua>on (Japan, Saudi Arabia) • University strategic planning and economic development (KAUST, Imam U, Princess Noura U) • Strategic technologies planning (Saudi Arabia, Korea) • Evalua>on of science, technology, and innova>on programs: Na>onal Science Founda>on, Na>onal Ins>tutes of Standards and Technology, Ohio • Training workshops in R&D management, program evalua>on, and interna>onal collabora>on (Korea, Saudi Arabia) • Design and implementa>on of technology development organiza>ons – na>onal labs, incubators, innova>on centers (Saudi Arabia)
Taxonomy Support for NCSES • SRI is a long‐>me support contractor for the Na>onal Center for Science & Engineering Sta>s>cs at the NSF (formerly the Division of Science Resource Sta>s>cs) • Engaged in 2009 for yet another project on classifica>ons of fields of science • Five (5) taxonomy reports since 2000 • One (1) SRS taxonomy team (2001/2002) • Two workshops (2006, 2008) • Dealing with perennial challenges of interdisciplinary research and taxonomy harmoniza>on across NCSES surveys and publica>ons • Working with Patrick Lambe of Straits Knowledge, Singapore & led by Jeri Mulrow, NCSES
Some Observa>ons So Far • Differences in taxonomies driven by data collec%on needs versus data analysis needs • What is it that we really want to classify? – Fields of research (what is going on in the lab) – Fields of study (NOT the same as disciplines) • Especially significant when studying Ph.D.’s and post‐docs – Fields of educa=on (roughly analogous to disciplines?) – Departments (organiza>onal units) – Occupa=ons – Fields of applica=on (technologies?) • Ojen >ed to markets/industries
One Illustra>on of the Problem Courtesy of Diana Hicks, Georgia Tech
One Approach: Faceted Taxonomies • Framework to classify in mul>ple dimensions – Each dimension may be a separate tree or list – Dimensions are orthogonal – mutually exclusive
Example: ANZSRC • Type of Ac>vity – Pure basic research – Strategic basic research – Applied research – Experimental Research • Field of Research – categorizes the methodology – 3 hierarchical levels – 22 Divisions, 157 Groups, 1238 Fields • Socio‐Economic Objec>ve – purpose or outcome – 5 Sectors; 17 Divisions, 119 Groups, 187 Objec>ves • Minor revisions every 5 years; major every 10
Current NCSES Ac>vi>es TAXONOMY MANAGEMENT TAXONOMY SCOPING R&D POLICY MANAGEMENT SYSTEM TAXONOMY • Provides a formal structure to • Provides a common • Helps us figure out what’s think about taxonomy environment to store, involved in developing a improvement and principles share, consult, link and taxonomy to be used as everyone can follow manage classifications and inter agency standard and vocabularies whether it’s worth the effort • Defines roles and responsibili>es, clarifies • Creates greater • Helps us clarify our expecta>ons opportunity for visibility into objectives and understand what we have our options • Guides decisionmaking on taxonomy improvement • Makes harmonisation less • Gives us a more confident towards greater manual and arduous and complete feel for harmonisa>on and currency what’s already out there • Makes it easier to show • Provides for feedback the ramifications of change • Helps us define the mechanism taxonomy strategy eg • Enables making reasoning, hierarchy, facets, principles, meanings of thesaurus terms, reasons for term exclusion explicit • Retains local control where needed
Opera>onalizing Taxonomy Management
Other Relevant Work • Exploratory project with STAR Metrics to develop taxonomies to work with topic modeling efforts in CHEM Directorate • Discussions of user needs for a mul>‐agency system for the Classifica>on of Research & Development Ac>vi>es (CORDA) • IARPA FUSE Project
Prospec>ve Areas of Inquiry • Do we really know what IS a science? • Technical approaches can eliminate SOME issues – Triangula>on to iden>fy and correct errors – Beper enforcement of consistent repor>ng policies – Unified format standards and data architectures – Increased use of machine analysis • Abandon the goal of a unified taxonomy – Create an environment which accommodates mul>ple compe>ng taxonomies – Leverage new technologies in text analysis, concept inference, Seman>c Web – Make taxonomies self‐organizing and self‐correc>ng – Requires compu>ng power, intensive design effort, and RESOURCES
Thank You Headquarters: Silicon Valley SRI Interna=onal 333 Ravenswood Avenue Menlo Park, CA 94025‐3493 650.859.2000 Washington, D.C. SRI Interna=onal 1100 Wilson Blvd., Suite 2800 Arlington, VA 22209‐3915 703.524.2053 Princeton, New Jersey SRI Interna=onal Sarnoff 201 Washington Road Princeton, NJ 08540 609.734.2553 Addi%onal U.S. and interna%onal loca%ons www.sri.com
Recommend
More recommend