DIMACS Challenges

DIMACS Challenges
Name	DIMACS Challenges
Established	1990s
Discipline	Computational science; discrete mathematics
Location	Rutgers University
Country	United States

DIMACS Challenges

The DIMACS Challenges were a sequence of coordinated problem-driven competitions and collaborative programs hosted by the Center for Discrete Mathematics and Theoretical Computer Science (DIMACS) that brought together researchers from Bell Labs, IBM Research, AT&T, Microsoft Research, Intel Corporation, Google Research, Courant Institute, Princeton University, Stanford University, Massachusetts Institute of Technology, Harvard University, Yale University, Columbia University, Cornell University, University of California, Berkeley, California Institute of Technology, University of Illinois Urbana–Champaign, University of Michigan, University of Toronto, University of Waterloo, McGill University, ETH Zurich, University of Cambridge, University of Oxford, École Polytechnique Fédérale de Lausanne, Max Planck Society, Los Alamos National Laboratory, Sandia National Laboratories, Lawrence Berkeley National Laboratory, National Science Foundation, Department of Energy (United States), Defense Advanced Research Projects Agency, National Institutes of Health, European Research Council, Royal Society, Simons Foundation, Heilbronn Institute for Mathematical Research, Clay Mathematics Institute, International Mathematical Union, Association for Computing Machinery, Society for Industrial and Applied Mathematics, Institute of Electrical and Electronics Engineers, SIAM Conference on Discrete Algorithms, International Symposium on Symbolic and Algebraic Computation, Conference on Neural Information Processing Systems, International Conference on Machine Learning, Annual ACM Symposium on Theory of Computing, IEEE Symposium on Foundations of Computer Science, ACM SIGMOD Conference, International Conference on Very Large Data Bases, USENIX Association, European Conference on Machine Learning and Principles and Practice of Knowledge Discovery in Databases, NeurIPS, AAAI Conference on Artificial Intelligence.

History

The program began in the 1990s as an initiative at Rutgers University supported by agencies such as the National Science Foundation and collaborators from industrial labs including Bell Labs and AT&T Research. Early efforts built on prior cooperative models exemplified by events at Los Alamos National Laboratory and coordination networks like the Mathematical Sciences Research Institute and the Institute for Advanced Study. Over successive rounds the Challenges incorporated input from stakeholders connected to IBM Research, Microsoft Research, Intel Corporation, Google Research, and national laboratories including Sandia National Laboratories and Lawrence Berkeley National Laboratory. Workshops and follow-up conferences were hosted at venues such as the Courant Institute, Princeton University, Stanford University, and international nodes including ETH Zurich and École Polytechnique Fédérale de Lausanne.

Objectives and Scope

The stated objectives aligned with priorities promoted by advisory bodies like the National Research Council and funding agencies such as the Department of Energy (United States): to define realistic benchmark problems, to stimulate algorithmic innovation at institutions including Harvard University and Massachusetts Institute of Technology, and to foster cross-disciplinary engagement among groups from University of California, Berkeley and University of Toronto. Scope covered industrial-scale applications linked to partners such as IBM Research and AT&T, foundational theory pursued by participants from Princeton University and Cornell University, and systems engineering interests reflected by involvement from Intel Corporation and Google Research. The program advocated reproducible evaluation consistent with standards used by organizations like Association for Computing Machinery and Institute of Electrical and Electronics Engineers.

Challenge Topics and Formats

Topics mirrored themes in conferences and institutions such as the Annual ACM Symposium on Theory of Computing, SIAM Conference on Discrete Algorithms, NeurIPS, International Conference on Machine Learning, ACM SIGMOD Conference, and included graph problems, network design, routing, optimization, satisfiability, and data mining. Formats combined timed competitions, long-term algorithmic benchmarks, and cooperative task forces similar to models employed by DARPA programs and consortium efforts seen at Max Planck Society centers. Datasets and problem instances were contributed by industrial partners like AT&T and Bell Labs as well as academic teams from University of Illinois Urbana–Champaign and University of Michigan, while evaluation criteria referenced norms from European Research Council funded projects and standards used by the Simons Foundation.

Organization and Participants

The Challenges were organized by DIMACS staff at Rutgers University in collaboration with program committees drawn from Princeton University, Cornell University, Stanford University, Massachusetts Institute of Technology, Harvard University, Yale University, Columbia University, University of California, Berkeley, and industrial research groups at Bell Labs, IBM Research, Microsoft Research, Intel Corporation, and Google Research. Funding and in-kind support came from agencies including the National Science Foundation, Department of Energy (United States), DARPA, and philanthropic bodies such as the Simons Foundation and the Clay Mathematics Institute. Participants ranged from graduate students and postdoctoral researchers affiliated with McGill University and University of Waterloo to senior researchers from Los Alamos National Laboratory and Sandia National Laboratories.

Impact and Outcomes

Outcomes included public benchmark repositories used by teams at IBM Research, Microsoft Research, and Google Research; spawning of algorithmic advances adopted in systems from Intel Corporation and cloud platforms maintained by Amazon Web Services and Microsoft Azure; and influence on curricula at academic departments such as Princeton University and Stanford University. Results informed follow-on programs at national labs like Los Alamos National Laboratory and policy discussions involving the National Science Foundation. Publications appeared in proceedings of the Association for Computing Machinery, SIAM, and conference outlets including NeurIPS and ACM SIGMOD Conference.

Selected Notable Challenges

Selected notable problem tracks engaged communities around graph clustering and community detection topics heavily cited by researchers at University of California, Berkeley, University of Cambridge, and ETH Zurich; maximum flow and multicommodity flow tracks with contributions from Princeton University and Cornell University; satisfiability and SAT solver comparison tracks influenced by advances at IBM Research and Microsoft Research; and VLSI routing and layout tracks relevant to Intel Corporation and Qualcomm. Other notable tracks intersected with machine learning and data mining research presented at NeurIPS and International Conference on Machine Learning and bioinformatics problem sets linked to projects funded by the National Institutes of Health and conducted at Harvard University and McGill University.

Category:Computer science competitions