Offset filtration

(Redirected from Offset Filtration)

The offset filtration (also called the "union-of-balls"[1] or "union-of-disks"[2] filtration) is a growing sequence of metric balls used to detect the size and scale of topological features of a data set. The offset filtration commonly arises in persistent homology and the field of topological data analysis. Utilizing a union of balls to approximate the shape of geometric objects was first suggested by Frosini in 1992 in the context of submanifolds of Euclidean space.[3] The construction was independently explored by Robins in 1998, and expanded to considering the collection of offsets indexed over a series of increasing scale parameters (i.e., a growing sequence of balls), in order to observe the stability of topological features with respect to attractors.[4] Homological persistence as introduced in these papers by Frosini and Robins was subsequently formalized by Edelsbrunner et al. in their seminal 2002 paper Topological Persistence and Simplification.[5] Since then, the offset filtration has become a primary example in the study of computational topology and data analysis.

The offset filtration at six scale parameters on a point cloud sampled from two circles of different sizes.

Definition

edit

Let   be a finite set in a metric space  , and for any   let   be the closed ball of radius   centered at  . Then the union   is known as the offset of   with respect to the parameter   (or simply the  -offset of  ).

By considering the collection of offsets over all   we get a family of spaces   where   whenever  . So   is a family of nested topological spaces indexed over  , which defines a filtration known as the offset filtration on  .[6]

Note that it is also possible to view the offset filtration as a functor   from the poset category of non-negative real numbers to the category of topological spaces and continuous maps.[7][8] There are some advantages to the categorical viewpoint, as explored by Bubenik and others.[9]

Properties

edit

A standard application of the nerve theorem shows that the union of balls has the same homotopy type as its nerve, since closed balls are convex and the intersection of convex sets is convex.[10] The nerve of the union of balls is also known as the Čech complex,[11] which is a subcomplex of the Vietoris-Rips complex.[12] Therefore the offset filtration is weakly equivalent to the Čech filtration (defined as the nerve of each offset across all scale parameters), so their homology groups are isomorphic.[13]

Although the Vietoris-Rips filtration is not identical to the Čech filtration in general, it is an approximation in a sense. In particular, for a set   we have a chain of inclusions   between the Rips and Čech complexes on   whenever  .[14] In general metric spaces, we have that   for all  , implying that the Rips and Cech filtrations are 2-interleaved with respect to the interleaving distance as introduced by Chazal et al. in 2009.[15][16]

It is a well-known result of Niyogi, Smale, and Weinberger that given a sufficiently dense random point cloud sample of a smooth submanifold in Euclidean space, the union of balls of a certain radius recovers the homology of the object via a deformation retraction of the Čech complex.[17]

The offset filtration is also known to be stable with respect to perturbations of the underlying data set. This follows from the fact that the offset filtration can be viewed as a sublevel-set filtration with respect to the distance function of the metric space. The stability of sublevel-set filtrations can be stated as follows: Given any two real-valued functions   on a topological space   such that for all  , the  -dimensional homology modules on the sublevel-set filtrations with respect to   are point-wise finite dimensional, we have   where   and   denote the bottleneck and sup-norm distances, respectively, and   denotes the  -dimensional persistent homology barcode.[18] While first stated in 2005, this sublevel stability result also follows directly from an algebraic stability property sometimes known as the "Isometry Theorem,"[9] which was proved in one direction in 2009,[16] and the other direction in 2011.[19][20]

A multiparameter extension of the offset filtration defined by considering points covered by multiple balls is given by the multicover bifiltration, and has also been an object of interest in persistent homology and computational geometry.[21][22]

References

edit
  1. ^ Adams, Henry; Moy, Michael (2021). "Topology Applied to Machine Learning: From Global to Local". Frontiers in Artificial Intelligence. 4: 2. doi:10.3389/frai.2021.668302. ISSN 2624-8212. PMC 8160457. PMID 34056580.
  2. ^ Edelsbrunner, Herbert (2014). A short course in computational geometry and topology. Cham. p. 35. ISBN 978-3-319-05957-0. OCLC 879343648.{{cite book}}: CS1 maint: location missing publisher (link)
  3. ^ Frosini, Patrizio (1992-02-01). Casasent, David P. (ed.). "Measuring shapes by size functions". Intelligent Robots and Computer Vision X: Algorithms and Techniques. 1607. Boston, MA: 122–133. Bibcode:1992SPIE.1607..122F. doi:10.1117/12.57059. S2CID 121295508.
  4. ^ Robins, Vanessa (1999-01-01). "Towards computing homology from approximations" (PDF). Topology Proceedings. 24: 503–532.
  5. ^ Edelsbrunner; Letscher; Zomorodian (2002). "Topological Persistence and Simplification". Discrete & Computational Geometry. 28 (4): 511–533. doi:10.1007/s00454-002-2885-2. ISSN 0179-5376.
  6. ^ Halperin, Dan; Kerber, Michael; Shaharabani, Doron (2015), Bansal, Nikhil; Finocchi, Irene (eds.), "The Offset Filtration of Convex Objects", Algorithms - ESA 2015, Lecture Notes in Computer Science, vol. 9294, Berlin, Heidelberg: Springer Berlin Heidelberg, pp. 705–716, arXiv:1407.6132, doi:10.1007/978-3-662-48350-3_59, ISBN 978-3-662-48349-7, S2CID 660889, retrieved 2023-02-25
  7. ^ Bauer, Ulrich; Kerber, Michael; Roll, Fabian; Rolle, Alexander (2023-02-16). "A unified view on the functorial nerve theorem and its variations". Expositiones Mathematicae. 41 (4): 8. arXiv:2203.03571. doi:10.1016/j.exmath.2023.04.005. S2CID 247291819.
  8. ^ Blumberg, Andrew J.; Lesnick, Michael (2022-10-17). "Stability of 2-Parameter Persistent Homology". Foundations of Computational Mathematics. arXiv:2010.09628. doi:10.1007/s10208-022-09576-6. ISSN 1615-3375. S2CID 224705357.
  9. ^ a b Bubenik, Peter; Scott, Jonathan A. (2014). "Categorification of Persistent Homology". Discrete & Computational Geometry. 51 (3): 600–627. arXiv:1205.3669. doi:10.1007/s00454-014-9573-x. ISSN 0179-5376. S2CID 254027425.
  10. ^ Edelsbrunner, Herbert (1993). "The union of balls and its dual shape". Proceedings of the ninth annual symposium on Computational geometry - SCG '93. San Diego, California, United States: ACM Press. pp. 218–231. doi:10.1145/160985.161139. ISBN 978-0-89791-582-3. S2CID 9599628.
  11. ^ Kim, Jisu; Shin, Jaehyeok; Chazal, Frédéric; Rinaldo, Alessandro; Wasserman, Larry (2020-05-12). "Homotopy Reconstruction via the Cech Complex and the Vietoris-Rips Complex". arXiv:1903.06955 [math.AT].
  12. ^ Edelsbrunner, Herbert (2010). Computational topology : an introduction. J. Harer. Providence, R.I.: American Mathematical Society. p. 61. ISBN 978-0-8218-4925-5. OCLC 427757156.
  13. ^ Chazal, Frédéric; Michel, Bertrand (2021). "An Introduction to Topological Data Analysis: Fundamental and Practical Aspects for Data Scientists". Frontiers in Artificial Intelligence. 4: 667963. doi:10.3389/frai.2021.667963. ISSN 2624-8212. PMC 8511823. PMID 34661095.
  14. ^ de Silva, Vin; Ghrist, Robert (2007-04-25). "Coverage in sensor networks via persistent homology". Algebraic & Geometric Topology. 7 (1): 339–358. doi:10.2140/agt.2007.7.339. ISSN 1472-2739.
  15. ^ Anai, Hirokazu; Chazal, Frédéric; Glisse, Marc; Ike, Yuichi; Inakoshi, Hiroya; Tinarrage, Raphaël; Umeda, Yuhei (2020-05-26). Topological Data Analysis. Abel Symposia. Vol. 15. arXiv:1811.04757. doi:10.1007/978-3-030-43408-3. ISBN 978-3-030-43407-6. S2CID 242491854.
  16. ^ a b Chazal, Frédéric; Cohen-Steiner, David; Glisse, Marc; Guibas, Leonidas J.; Oudot, Steve Y. (2009-06-08). "Proximity of persistence modules and their diagrams". Proceedings of the twenty-fifth annual symposium on Computational geometry. Aarhus Denmark: ACM. pp. 237–246. doi:10.1145/1542362.1542407. ISBN 978-1-60558-501-7. S2CID 840484.
  17. ^ Niyogi, Partha; Smale, Stephen; Weinberger, Shmuel (2008). "Finding the Homology of Submanifolds with High Confidence from Random Samples". Discrete & Computational Geometry. 39 (1–3): 419–441. doi:10.1007/s00454-008-9053-2. ISSN 0179-5376. S2CID 1788129.
  18. ^ Cohen-Steiner, David; Edelsbrunner, Herbert; Harer, John (2007). "Stability of Persistence Diagrams". Discrete & Computational Geometry. 37 (1): 103–120. doi:10.1007/s00454-006-1276-5. ISSN 0179-5376.
  19. ^ Lesnick, Michael (2015). "The Theory of the Interleaving Distance on Multidimensional Persistence Modules". Foundations of Computational Mathematics. 15 (3): 613–650. arXiv:1106.5305. doi:10.1007/s10208-015-9255-y. ISSN 1615-3375. S2CID 254158297.
  20. ^ Lesnick, Michael (2023). "Lecture notes for AMAT 840: Multiparameter Persistence" (PDF). University at Albany, SUNY.
  21. ^ Corbet, René; Kerber, Michael; Lesnick, Michael; Osang, Georg (2023-02-20). "Computing the Multicover Bifiltration". Discrete & Computational Geometry. 70 (2): 376–405. arXiv:2103.07823. doi:10.1007/s00454-022-00476-8. ISSN 0179-5376. PMC 10423148. PMID 37581017.
  22. ^ Edelsbrunner, Herbert; Osang, Georg (2021). "The Multi-Cover Persistence of Euclidean Balls". Discrete & Computational Geometry. 65 (4): 1296–1313. doi:10.1007/s00454-021-00281-9. ISSN 0179-5376. PMC 8550220. PMID 34720303.