Generalized Metric Repair on Graphs

Authors Chenglin Fan, Anna C. Gilbert, Benjamin Raichel, Rishi Sonthalia, Gregory Van Buskirk

Thumbnail PDF


  • Filesize: 0.84 MB
  • 22 pages

Document Identifiers

Author Details

Chenglin Fan
  • Department of Computer Science, University of Texas at Dallas, Richardson, TX 75080, USA
Anna C. Gilbert
  • Department of Mathematics, University of Michigan, Ann Arbor, MI 48109, USA
Benjamin Raichel
  • Department of Computer Science, University of Texas at Dallas, Richardson, TX 75080, USA
Rishi Sonthalia
  • Department of Mathematics, University of Michigan, Ann Arbor, MI 48109, USA
Gregory Van Buskirk
  • Department of Computer Science, University of Texas at Dallas, Richardson, TX 75080, USA

Cite AsGet BibTex

Chenglin Fan, Anna C. Gilbert, Benjamin Raichel, Rishi Sonthalia, and Gregory Van Buskirk. Generalized Metric Repair on Graphs. In 17th Scandinavian Symposium and Workshops on Algorithm Theory (SWAT 2020). Leibniz International Proceedings in Informatics (LIPIcs), Volume 162, pp. 25:1-25:22, Schloss Dagstuhl – Leibniz-Zentrum für Informatik (2020)


Many modern data analysis algorithms either assume or are considerably more efficient if the distances between the data points satisfy a metric. However, as real data sets are noisy, they often do not possess this fundamental property. For this reason, Gilbert and Jain [A. Gilbert and L. Jain, 2017] and Fan et al. [C. Fan et al., 2018] introduced the closely related sparse metric repair and metric violation distance problems. Given a matrix, representing all distances, the goal is to repair as few entries as possible to ensure they satisfy a metric. This problem was shown to be APX-hard, and an O(OPT^{1/3})-approximation was given, where OPT is the optimal solution size. In this paper, we generalize the problem, by describing distances by a possibly incomplete positively weighted graph, where again our goal is to find the smallest number of weight modifications so that they satisfy a metric. This natural generalization is more flexible as it takes into account different relationships among the data points. We demonstrate the inherent combinatorial structure of the problem, and give an approximation-preserving reduction from MULTICUT, which is hard to approximate within any constant factor assuming UGC. Conversely, we show that for any fixed constant ς, for the large class of ς-chordal graphs, the problem is fixed parameter tractable, answering an open question from previous work. Call a cycle broken if it contains an edge whose weight is larger than the sum of all its other edges, and call the amount of this difference its deficit. We present approximation algorithms, one depending on the maximum number of edges in a broken cycle, and one depending on the number of distinct deficit values, both quantities which may naturally be small. Finally, we give improved analysis of previous algorithms for complete graphs.

Subject Classification

ACM Subject Classification
  • Theory of computation → Computational geometry
  • Approximation
  • FPT
  • Hardness
  • Metric Spaces


  • Access Statistics
  • Total Accesses (updated on a weekly basis)
    PDF Downloads


  1. N. Alon and S. Gutner. Balanced families of perfect hash functions and their applications. ACM Trans. Algorithms, 6(3):54:1-54:12, 2010. Google Scholar
  2. S. Baraty, D. Simovici, and C. Zara. The impact of triangular inequality violations on medoid-based clustering. In Foundations of Intelligent Systems, pages 280-289. Springer, 2011. Google Scholar
  3. M. Belkin and P. Niyogi. Laplacian eigenmaps for dimensionality reduction and data representation. Neural Comput., 15(6):1373-1396, June 2003. URL:
  4. C. Brand, H. Dell, and T. Husfeldt. Extensor-coding. In Symposium on Theory of Computing (STOC), pages 151-164, 2018. Google Scholar
  5. J. Brickell, I Dhillon, S. Sra, and J. Tropp. The metric nearness problem. SIAM Journal on Matrix Analysis and Applications, 30(1):375-396, 2008. Google Scholar
  6. E. Candès and B. Recht. Exact matrix completion via convex optimization. Commun. ACM, 55(6):111-119, June 2012. Google Scholar
  7. L. Sunil Chandran, Vadim V. Lozin, and C. R. Subramanian. Graphs of low chordality. Discrete Mathematics and Theoretical Computer Science, 7:25-36, 2005. Google Scholar
  8. S. Chawla, R. Krauthgamer, R. Kumar, Y. Rabani, and D. Sivakumar. On the hardness of approximating multicut and sparsest-cut. Computational Complexity, 15(2):94-114, 2006. Google Scholar
  9. C. Fan, B. Raichel, and G. Van Buskirk. Metric violation distance: Hardness and approximation. In Symposium on Discrete Algorithms (SODA), pages 196-209, 2018. Google Scholar
  10. A. Gilbert and L. Jain. If it ain't broke, don't fix it: Sparse metric repair. ArXiv, 2017. Google Scholar
  11. S. Khot. On the power of unique 2-prover 1-round games. In Proceedings on 34th Annual ACM Symposium on Theory of Computing (STOC), pages 767-775, 2002. Google Scholar
  12. E. Lee. Improved hardness for cut, interdiction, and firefighter problems. In 44th International Colloquium on Automata, Languages, and Programming (ICALP), pages 92:1-92:14, 2017. Google Scholar
  13. S. Roweis and L. Saul. Nonlinear dimensionality reduction by locally linear embedding. Science, 290(5500):2323-2326, 2000. URL:
  14. A. Sidiropoulos, D. Wang, and Y. Wang. Metric embeddings with outliers. In Proc. Twenty-Eighth Annual ACM-SIAM Symposium on Discrete Algorithms (SODA), pages 670-689, 2017. Google Scholar
  15. J. Tenenbaum, V. Silva, and J. Langford. A global geometric framework for nonlinear dimensionality reduction. Science, 290(5500):2319-2323, 2000. URL:
  16. L. Valiant. The complexity of enumeration and reliability problems. SIAM J. Comput., 8(3):410-421, 1979. Google Scholar
  17. F. Wang and J. Sun. Survey on distance metric learning and dimensionality reduction in data mining. Data Mining and Knowledge Discovery, 29(2):534-564, March 2015. Google Scholar