Theoretical Computer Science at NSF: Past, Present, and Future

1
Theoretical Computer Science at NSF Past,
Present, and Future

• Michael Foster
• Division Director
• CISE/CCF
• (with D. Du and S. Mahaney)

2
But First

• Some Organizational Context

3
Funding Outlook

• NSF funds available to support computing have
  nearly doubled in the past five years
• However, proposals have almost tripled
• From less than one per year per CS faculty member
  to more than one per year.
• CISE budget outlook for near future
• slow growth or small decline likely
• transition of ITR funds into core programs

4
Whats Theory For?

• Hard foundational questions
• Linkages between disparate fields
• Incubator

5
Theory PI Trajectories
Other NSF
Career
8, 9
ITR
2, 3, 7, 8, 9
5
Distance from Core
10
3, 4, 5
1,5,6, 10, 11
7
4
1,2,3,4, 6
TOC
1,2, 6, 7, 8, 9, 10, 12
Risk
6
Strategies

• Division
• Encourage research at the boundaries
• Some success at boundary of Communications and
  TOC
• Gradually increase cluster reserves to 60
• Theory Program
• Maintain strong performers in complexity
• Widespread support rather than large awards
• Support new investigators
• Look for new results in cryptography
• Narrow system-theory gap in algorithms
• Theoretical foundation for good heuristics
• Approximation algorithms
• Applied theory
• Co-fund with networks, biology, etc.

7
Program Directors Sought

• Communication Theory
• Theory of Computing
• Computer Architecture
• Emerging Models and Technologies
• Interdisciplinary capability
• Across cluster, division, NSF, and globally

8
Contact

• Vacancy announcements appear on www.nsf.gov
• Meanwhile contact
• Michael Foster
• Division Director
• Computing Communication Foundations
• National Science Foundation
• 4201 Wilson Boulevard
• Arlington, VA 22230
• 703-292-8910
• mfoster_at_nsf.gov

9
Backups
10
NSF Organization
National Science
Office of the Director
Board
Directorate for Biological
Directorate for Mathematical
Sciences
Physical Sciences
Directorate for Computer
Directorate for Social, Behavioral
Economic Sciences
Information Science and Eng.
Directorate for Education
Directorate for Geosciences
Human Resources
Directorate for Engineering
Office of Polar Programs
Administrative Offices
11
New CISE Organization
Office
of the
Assistant
Director
Computing and
Computer and
Information and
Shared
Communication
Network
Intelligent
Cyberinfrastructure
Foundations
Systems
Systems
(SCI)
(CCF)
(CNS)
(IIS)
12
Clusters within Divisions

• Comprehensive activity in a coherent area of
  research and education
• Teams of Program Officers and Staff working
  closely with the community
• Initially groups of existing programs
• Eventually one program per cluster

13
Computing and Communication Foundations (CCF)

• Theoretical Foundations
• Computer science theory numerical computing
  computational algebra and geometry signal
  processing and communication
• Foundations of Computing Processes and Artifacts
• Software engineering software tools for HPC
  programming languages compilers computer
  architecture graphics and visualization
• Emerging Models and Technologies for Computation
• Computational biology quantum computing
  nano-scale computing biologically inspired
  computing

14
CCF Cluster Competitions

• FY 2004
• Responsible for about 1900 competitive proposals
• Heavy mortgages and commitments
• Competitive proposal success rate 18
• FY 2005 and FY 2006 One solicitation per cluster
• CAREER competition almost complete (16 success
  in 2005)
• Theoretical Foundations January 2005 due date,
  received about 440 proposals. 2006 competition
  TBD
• Emerging Models and Technologies for Computation
  February 2005 due date, received 150
  proposals. 2006 competition TBD
• Foundations of Computing Processes and Artifacts
  Due Date June 20, 2005
• FY 2007 Fall or winter deadlines for all three
  clusters

15
Example TOC Projects

• Eva Tardos Approximation Algorithms and Network
  Games
• Design of games that model contention in
  distributed networks
• Teng and Spielman ITR Smoothed Analysis of
  Algorithms
• Average case analysis using perturbed inputs
• Shows why Simplex works in practice
• Wigderson ITR Computational Complexity Theory
  2003
• Group support for Princeton IAS
• Amit Chakrabarti CAREER
• Quantify information cost and address
• Multiparty set disjointness and variants,
• Relative power of quantum and classical
  communication models
• Other comunication complexity problems

16
Current TOC Portfolio

• Core theory
• Computational complexity, other lower bounds
• Fundamental Algorithms
• Transcend application domains combinatorial,
  approximation, parallel, on-line, graph
  algorithms,
• Application-specific theory
• Includes molecular biology, networks, linguistics

.
17
Theory Support and Publication

• A small sample
• first 16 papers of STOC 2004 proceedings
• NSF Funding status as of Spring 2004
• Out of 12 eligible authors, 10 have NSF support
• 37 distinct authors, 2 repeats
• Not eligible for NSF funding
• 15 not from US
• 3 from US Industry
• 6 US students (1 NSF Grad Research Fellowship)
• 1 US Postdoc
• Eligible for NSF funding
• 2 not funded (1 funded after Spring 2004)
• 10 funded by NSF
• 6 from Theory of Computing program
• 5 ITR awards
• 1 MPS/Math
• 3 other CISE

18
CISE BudgetFY05 Current Plan (M)
19
Historical TOC Statistics
Sources EIS, 2002 COV report, PD data
20
Theory Career Awards (00-04)
X
Other NSF
Career
X
X
X
ITR
X
X
X
X
X
X
Distance from Core
X
X
X
X
X
X
X
X
X
X
X
X
X
TOC
X
X
X
X
X
X
X
Risk
21
Ideal Support for Theory
Other NSF
Career
ITR
Distance from Core
TOC
Risk
22
Why This Session

• Update on a core CISE area
• Continuing CISE reorganization
• Impact of NSF/CISE budgets
• Evidence of community dissatisfaction
• Advice requested

23
Advice, please

• Division
• Modify speed of clustering?
• Recluster?
• Theory program
• Maintain strong performers for hard foundational
  questions?
• Continue strategy of diverse awards?
• Encourage impure theory to link fields?
• Encourage co-funding to incubate?