CENTER Carnegie Mellon UniversityCarnegie Mellon Computer Science DepartmentSchool of Computer Science
Problem-Oriented Explorations
Related Activities
Outreach Roadshow

The Center's research is structured as an overlapping sequence of PROBEs (PROBlem-oriented Explorations). Each PROBE focuses on a specific problem domain, and lasts for approximately one year. Its purpose is to find algorithmic solutions and appropriate mathematical formulations for that problem domain, and thus involves both domain experts as well algorithm designers.

The PROBE topics are selected to have some common themes, so that ideas are shared among the PROBEs. Each PROBE provides funding for faculty, students, and possibly a postdoctoral fellow, as well as travel funding for the workshops. To maximize pay-off, PROBEs seek broadly-applicable solutions to domain problems, and not specialized solutions for specific scenarios nor detailed solutions for the purpose of commercialization (i.e., research, not product development).


Flexible Network Design
Spectral Graph Theory and Applications
Algorithms for SNP and Haplotype Analysis
Algorithms in Economics
Computer-Human Authentication with Applications to AI
Dynamic Algorithms and Applications
Graph Cuts for Vision and Data Analysis
Integrated Logistics
Meshing, Theory and Applications
Privacy In D.A.T.A.
Scheduling Your Network Connections
Utilizing the Power of Human Cycles
Web Structure and Algorithms

PROBE selection

The PROBE topics are selected by the Scientific Advisory Committee. We encourage people to submit proposals. A proposal should be from 2-3 pages and contain the following:

  • An overview of the domain
  • The current state-of-the-art in algorithms applied to the domain.
  • The set of algorithms/problems that the PROBE will explore
  • A list of PROBE organizers (from 2-4).

The PROBEs will be selected based on the following three criteria:

  • Potential impact
  • The idea of a PROBE is not to explore something known to work, but rather to explore something which might work and can have significant impact.
  • Cross-disciplinary participation. The suggested participants must include members from other institutions and must include both domain experts and algorithm designers. Furthermore, it must also contain some people who can help bridge the gap. Ties to other PROBEs. There should be connections between the PROBE and other existing or potential future PROBEs.
  • PROBE budget The budget can vary from PROBE to PROBE and we are always looking for some matching support from other sources. Assuming no external funding, a PROBE will cover about 3.5 faculty summer salaries, 2.5 students, .5 postdocs, $10K travel, and $15K "workshop" costs. This, however, can be balanced in various ways.

PROBE Schedule

In general The PROBEs run in a synchronized schedule, from June-to-June. Each PROBE is punctuated by three workshop-style meetings, each lasting from one to three days. In the first meeting (in June), the participants gather for tutorials on the topic, and to initiate a discussion between the domain experts and algorithm designers. This is expected to jump start a productive summer of collaborative research. The second meeting (in October) allows the same participants to discuss progress to date, including detailed proposals of algorithms and techniques to apply within the domain, and new theoretical problems to evaluate. These algorithms, techniques and problems are explored and evaluated in further detail between the second and third meetings. The third meeting (in June) presents results of the research. While the first and third meetings are open to anyone,the second only involves the main participants.

PROBE results

Each PROBE produces several concrete results, including published papers, running code, pages for our web repository, data sets, and case studies for the education component. In addition to these concrete results, there are many other benefits of each PROBE. These can include new interesting theoretical problems, new ways to characterize data, new ways of formulating and viewing problems within the domains, and the transfer of knowledge and techniques across domains. Also, as the term PROBE and its expansion imply, it is our expectation that most of the PROBEs will lead to longer-term studies of the respective areas. These longer term studies will not be directly funded as part of the Center, but infrastructure within the Center will help find funding for these follow-on studies, especially from industry.

Each PROBE will assign a student scribe and a student code manager. The Scribe will be responsible for documenting the problems and algorithms that were explored. This will be used to help the organizers write the case study document. The Code Manager will be responsible for collecting data sets and algorithms developed as part of the PROBE.

PROBEs will be expected to produce the following concrete results: (1) tutorials covering the domain, and covering relevant algorithms in the domain, (2) a case study document, (3) code and data, (4) a set of web pages on the topic, (5) publishable papers, and (6) a course on the topic.


This material is based upon work supported by National Science Foundation under Grant No. 0122581.
Any opinions, findings, and conclusions or recommendations expressed in this material are those of the author(s) and do not necessarily reflect the views of the
National Science Foundation