A proposal for FY13 UTK Organized Research Unit Funding

Presented by: Dr. Lynne Parker, Director

March 5, 2012

The Vision for CISML

Develop interdisciplinary theory and practice of intelligent systems and machine learning technologies

Enable cross-fertilization of ideas from several individual disciplines Attract increased external funding involving multiple faculty Connect industry with curricula and research Help UTK reach its Top 25 goal, by cultivating our established strengths in

intelligent systems and machine learning Attract more highly qualified students Integrate curricular content and emphasize interdisciplinary study

2

CISML Organization

Dr. Lynne ParkerDirector

Dr. Michael Berry Assoc. Director

Mr. Scott WellsProgram Manager

Approved as a formal UTK Center in October, 2011

Organizational and Management Chart

Executive BoardContributing Faculty

Affiliates

DirectorLynne Parker (EECS)

Associate DirectorMike Berry (EECS)

Faculty Affiliates

Industrial Affiliates

National Lab Affiliates

Itamar Arel (EECS)Michael W. Berry (EECS)Ham Bozdogan (SOMS)Daniela Corbetta (Psych)Wes Hines (NE)Bruce MacLennan (EECS)Lynne Parker (EECS)Hairong Qi (EECS)Shih-Lung Shaw (Geog)

Wenjun Zhou (SOMS)

Catalyst Repository Systems

Huron LegalLink AnalyticsM-CAM

(Others, TBD)

Jacob BarhenJohn GoodallTom PotokVladimir ProtopopescuBrian WorleySonghua Xu

(Others, TBD)

Internal Advisory BoardBill Dunne, Engineering

Chris Boake, Arts & SciencesTom Ladd, Business

COE Associate Dean for Research and Technology (Bill Dunne)

Feb. 2012

Program ManagerScott Wells

CISML UTK Faculty – From 3 Colleges, 4 Depts.

4

College of Engineering:

CISML Director: Dr. Lynne Parker, Electrical Engineering and Computer Science (EECS)

Dr. Itamar Arel, EECS Dr. Michael Berry, EECS Dr. J. Wes Hines, Nuclear

Engineering Dr. Bruce MacLennan,

EECS Dr. Hairong Qi, EECS

College of Business Administration

Dr. Ham Bozdogan, Statistics, Operations, and Mgmt. Sci. (SOMS)

Dr. Wenjun Zhou, SOMs

College of Arts and Sciences

Dr. Daniela Corbetta, Psychology

Dr. Shih-Lung Shaw, Geography

CISML Nat’l Lab Affiliates – from 2 Divisions, 4 groups

5

Computer Science and Mathematics Division:

Dr. Jacob Barhen, Complex Systems Group

Dr. Tom Potok, Applied Software Engineering Group

Computational Science and Engineering Division

Dr. Brian Worley, CSE Director

Dr. Vladimir Protopopescu, CSE Chief Scientist

Dr. John Goodall, Cyber Security and Information Infrastructure Research Group

Dr. Songhua Xu, Early Career Biomedical Research

CISML Industry Affiliates

6

More industry affiliates being recruited …

Each industry affiliate contributes $5K annually

In return, their benefits are:– Access to undergrad and grad students for

internships, employment

– Collaborative research with CISML

– Access to all public domain software developed, with opportunities for licensing

– Access to faculty and student research publications

– Display of corporate logo on website

– Participation in Industrial Affiliate workshop

– Recognition as CISML Industrial Affiliate

External Funding Opportunities

Important sponsors with broad open BAAs relevant to CISML include AFOSR, ARO, DARPA, DOE, IARPA, NIH, NSF, ONR, etc.

Other Industries currently engaged with CISML faculty (but not yet Affiliates) include: Pilot Travel Centers, Voices Heard Media, Capital One, Computable Genomix, SAS, Luxottica Retail, Document Solutions, Inc., and Lockheed Martin Advanced Technology Laboratories

7

Opportunities are Numerous and Significant:Many potential applications Many funding sponsors

Example applications:

Energy applications– E.g., Building energy prediction

Environmental monitoring– E.g., prediction of volcanic eruptions

Geographic Information Systems– E. g., people tracking

Medical diagnosis– E.g., Breast cancer detection, diagnostic

imaging, detection of cause of heart attack

Text and data mining– E.g., Electronic discovery

Cognitive computing and robotic learning– E.g., Using infant perceptual-motor

learning Reliability and prognostics

– E.g., in nuclear reactors, multi-robot systems

Intelligent transportation systems– E.g., automatic detection of incidents,

maximizing flow8

Key Objective of CISML: Leverage Research Synergies to Pursue Multi-Collaborator Funding

Strategy:

– Identify unique synergies amongst CISML Faculty, National Lab, and Industry Affiliates

• Through extensive discussions, CISML seminars, cross-fertilization of ideas

– Leverage synergies to pursue new directions for multi-collaborator, multi-disciplinary research

– Explore and pursue opportunities to participate in UTK, State, and National Initiatives

• E.g., in Energy/Power, national security and non-proliferation, manufacturing, etc.

– Explore and pursue opportunities for Center-level funding

• E.g., with NSF, DOE, etc.

12

Building CISML Synergies from Existing Competencies

CISML Affiliates have broad expertise in Intelligent Systems and Machine Learning:

– Reinforcement learning, deep machine learning (Arel, Parker)– Text/data mining and knowledge discovery (Berry, Bozdogan, Goodall,

Parker, Potok, Worley, Xu, Zhou)– Human infant perceptual and motor learning (Corbetta)– Cognitive learning (Arel, Corbetta)– Pattern recognition (Barhen, Berry, Hines, Parker, Qi)– Prognostics and diagnostics (Hines, Parker)– Embodied intelligence (Arel, Corbetta, MacLennan, Parker)– Collaborative/Cooperative/Distributed systems (Parker, Potok,

Protopopescu, Qi)– Remote sensing (Barhen, Parker)– Biologically-inspired intelligence (Arel, MacLennan, Parker, Potok)

13

Active CISML Research Synergies

Identified thus far:

– Using psychological studies of human infants’ grasping/manipulation learning to inform how to build smarter robotic systems• CISML Affiliates Involved: Arel, Corbetta, MacLennan, Parker• Preliminary research underway

– Analyzing human infant grasping trajectories for identifying motion patterns• CISML Affiliates Involved: Corbetta, Parker• Preliminary research underway

– Using models of visual attention built from human infant studies to develop high-performing computational models• CISML Affiliates Involved: Arel, Corbetta• Preliminary research underway

– Using Microsoft’s Kinect motion sensing device to capture intensity and depth information to detect and create models for 3D objects • CISML Affiliates Involved: Parker, Arel• Preliminary research underway

14

High Priority: Continue to define synergistic opportunities

Active CISML Research Synergies

– Using locomotion tracking of fish larvae to predict biological and environmental behavior changes induced by drugs and toxins• CISML Affiliates Involved: Parker, MABE dept. faculty, • Preliminary research underway

– Using algorithms to automatically catalog document objects and creates data types for ease of computational analysis• CISML Affiliates Involved: Berry, Catalyst Repository Systems• Preliminary research underway

– Using novel technologies, including eye-tracking, for improving the search of relevant online literature based on the segmentation of image, text, and audio data• CISML Affiliates Involved: Berry, Xu• Preliminary research underway

15

Active CISML Research Synergies

16

– Creating software (called PolyLens) to perform in-depth visual analysis of genome-scale polymorphism data• CISML Affiliates Involved: Berry• Preliminary research underway

– Analyzing and increasing functionality of the Cyton Alpha 7D 1G arm via enhancement of the Robot Operating System (ROS) architecture• CISML Affiliates Involved: Parker, Arel• Preliminary research underway

– Using Artificial Neural Network (ANN)-based classifiers to identify disturbances in electric power systems• CISML Affiliates Involved: Parker• Preliminary research underway

CISML Accomplishments since last review

Established a permanent/comprehensive web presence (http://cisml.utk.edu/)

Continued bi-weekly research seminar series; 18 seminars held to date (presentations available on website)

Produced center Bylaws (available on website)

Established a collaborative relationship with the Center for Brain-Like Computing and Machine Intelligence in Shanghai, China

18

CISML Accomplishments since last review

Developed annual Industry Affiliate Workshop (1st one held in April 2011)

Established a Distinguished Research Seminar Series co-sponsored by ORNL

Added three new Industry Affiliates

Added two new Faculty Affiliates

19

Wenjun ZhouStatistics, OperationsAnd Management Science

Shih-Lung ShawGeography

CISML Accomplishments since last review

Six proposal submissions

Produced multiple internships with existing industry affiliates

Established a new seminar course for PhD students in machine learning

Established a more unified presence with move into the new Min H. Kao building

20

CISML Plans and Goals for Year 3 (FY13)

Identification of new multi-investigator research synergies– Goal: 5+ new synergies identified

Increased number of multi-investigator proposals submitted– Goal: 6+ new proposals

Increased number of multi-investigator publications and presentations– Goal: 5+ papers/presentations

Increased interaction with potential sponsors– Goal: 4 visits and discussions with potential sponsors

Provide seed funds to CISML faculty for promising new research

Continuation of Distinguished Research CISML Seminar Series– Goal: 2 distinguished speakers

Increased number of Industrial Affiliate sponsors– Goal: 2+ new Industrial Affiliates

Increased number National Laboratory Affiliates– Goal: 1 new researcher

21

New CISML Activities for FY13

Support travel for CISML faculty to visit potential research sponsors, research program planning workshops, etc.– Numerous federal funding sponsors, numerous faculty

• Highly interdisciplinary nature and diversity of our research results in large pool of programs and program managers, etc.

• Likewise, many workshops are offered

Support seed money research funds for CISML faculty to pursue preliminary investigations– Funds will be competitive

• Require identification of specific funding opportunities to be pursued, expected publication venue(s), and expected benefit to CISML

– Funds will primarily support student stipends– Faculty will be required to submit developed proposals through CISML

22

24

Emerging Synergies for FY13

Grasping eDiscovery momentum with existing and prospective industry affiliates and ORNL

– Build a coalition of CISML affiliates around the opportunities presented for developing new and novel algorithms and software for eDiscovery.

– Definition: Electronic discovery (eDiscovery) is the process of collecting and analyzing electronic documents to determine their relevance to a legal matter.

– Participants

– Faculty affiliates: Berry, Parker, Zhou – ORNL affiliates: Goodall, Potok, Worley, Xu– Industry affiliates: Huron Legal, M-CAM, Catalyst Repository Systems

Emerging Synergies for FY13

– Vision and Strategy– Capitalize on new opportunities that have arisen to establish and continue

developing learning-based technologies that assist software developers, service providers, and legal teams in efficiently and accurately identifying the most relevant documents for litigation.

– Pursue center level funding via NSF’s Industry/University Cooperative Research Centers Program (I/UCRC)

• Funding Amount - $10,000,000• Centers funded – 2 - 8

From NSF Solicitation “The Industry/University Cooperative Research Centers (I/UCRC) program develops long-term partnerships among industry, academe, and government. An I/UCRC contributes to the Nation's research infrastructure base and enhances the intellectual capacity of the engineering and science workforce through the integration of research and education.”

25

Budget and Requested Funds for CISML

27

Budgeted Expenditures FY 2011 FY 2012 FY 2013 FY 2014CISML Program Manager (salary & benefits) $50,029 $101,931 $103,510 $106,678IT support (web, publications database, etc.) $6,100 $0 $0 $0Setup CISML main office (computer/printer/maint.) $4,035 $0 $0 $0Travel to potential sponsors $1,300 $0 $10,000 $10,000Seed research projects $0 $0 $7,500 $15,000Distinguished seminar series $0 $1750 $3,500 $5,000Other seminar and workshop expenses $140 $300 $300 $400Industrial Affiliates annual meeting $0 $500 $1,000 $1,500CISML Telephone $97 $375 $375 $375UT Extension (Ag Research) subcontract (program./design) $0 $25,683 $11,880 $0

Dr. Michael Berry (1 month summer) $0 $0 $16,064 $0Misc. operating expenses (e.g., supplies) $0 $300 $300 $400

Total budgeted expenditures for CISML $61,701 $130,839 $154,429 $139,353Sources of funding CISML faculty and dept. contributions, F&A recoveries $31,500 $20,000 $23,500 $45,500Industrial Affiliates $15,028 $30,000 $40,000 $50,000UT Extension subcontract $0 $31,050 $31,050 $15,525Science Alliance $15,000 $0 $0 $0Office of Research ORU Funding (requests for FY13-14) $30,000 $30,000 $55,000 $35,000

Total new income $91,528 $111,050 $149,550 $146,025Carryover from previous FY (industry, etc.) $0 $29,827 $16,038 $11,159

Grand total funds available $91,528 $140,877 $165,588 $157,184

Expected Returns are Significant

Increased Funding: CISML will enable UTK faculty to attract significant collaborative funding that otherwise would not be possible.

Innovative Research: CISML will develop new research directions enabled by cross-fertilization of ideas, to achieve multi-disciplinary, collaborative synergies

International Recognition: CISML will be recognized as a national and international leader in intelligent systems and machine learning

Higher Caliber Students: UTK will be better able to recruit high-caliber undergraduate and graduate students and postdocs

28

Thank you for your consideration!

29

Lynne ParkerItamar Arel Michael Berry Ham

Bozdogan Daniela Corbetta

Wes HinesShih-Lung Shaw

Bruce MacLennan

Hairong Qi

CISML Faculty

Wenjun Zhou

Backup Slides

30

Who is our “competition”?

Carnegie Mellon University, Machine Learning Department– 24 core faculty, 27 affiliated faculty, ~20 related faculty– Highly interdisciplinary

UC Berkeley, Center for Intelligent Systems– 26 faculty & research staff– Highly interdisciplinary

UC Irvine, Center for Machine Learning and Intelligent Systems – 30 faculty & research staff– Highly interdisciplinary

George Washington Univ., Center for Intelligent Systems Research – 12 faculty & research staff– Emphasis is on Intelligent Transportation Systems

Vanderbilt, Center for Intelligent Systems– 7 faculty & research staff– Emphasis is on robotics

University of Idaho, Center for Intelligent Systems Research– 6 faculty and research staff– Emphasis is primarily control

31

By joining efforts, UTK’s CISML can become a highly competitive research center

Achieved All Year 1 Goals – As Presented Last Year

Integrate and increase research activity– Metric: # of proposals submitted

• Accomplishment: 4 multi-CISML-investigator proposals submitted

Initiate research discussion groups– Metric: creation of regular discussion group meeting

• Accomplishment: Bi-weekly seminar series begun; 9 held to date

Engage in CISML-relevant external workshops and symposia– Metric: faculty actively involved

• Accomplishment: Significant faculty involvement in several relevant technical meetings

Integrate and enhance curricular offerings– Metric: development of showcase curriculum

• Accomplishment: Underway; will be completed by end of year 1

Establish a web presence– Metric: creation of the website

• Accomplishment: Done (http://cisml.utk.edu)

32

Also Achieved More Ambitious Year 1 Goals – As Evolved During Year

We accomplished all original Year 1 goals

Additionally, we expanded our Year 1 goals to include:– Create Center infrastructure that will facilitate successful collaborations

• Accomplishments: – Hired CISML program manager– Identified CISML office space– Applied for, and was approved, as official UTK Center– Identified personnel to handle CISML finance reporting

– Identify research synergies among CISML Affiliates• Accomplishments:

– Creation of regular research discussion group– Identification of 5 synergies thus far

– Pursue funding opportunities that leverage research synergies• Accomplishments:

– 4 proposals submitted thus far

33