DOCSLIB.ORG

Nash Embedding Theorem - Wikipedia, the Free Encyclopedia Page 1 of 3

Total Page：16

File Type：pdf, Size：1020Kb

Download full-text PDF Read full-text
Nash Embedding Theorem - Wikipedia, the Free Encyclopedia Page 1 of 3 Nash embedding theorem - Wikipedia, the free encyclopedia Page 1 of 3 Nash embedding theorem From Wikipedia, the free encyclopedia The Nash embedding theorems (or imbedding theorems ), named after John Forbes Nash, state that every Riemannian manifold can be isometrically embedded into some Euclidean space. Isometric means preserving the length of every path. For instance, bending without stretching or tearing a page of paper gives an isometric embedding of the page into Euclidean space because curves drawn on the page retain the same arclength however the page is bent. The first theorem is for continuously differentiable ( C1) embeddings and the second for analytic embeddings or embeddings that are smooth of class Ck, 3 ≤ k ≤ ∞. These two theorems are very different from each other; the first one has a very simple proof and is very counterintuitive, while the proof of the second one is very technical but the result is not at all surprising. The C1 theorem was published in 1954, the Ck-theorem in 1956. The real analytic theorem was first treated by Nash in 1966; his argument was simplified considerably by Greene & Jacobowitz (1971). (A local version of this result was proved by Elie Cartan and Maurice Janet in the 1920s.) In the real analytic case, the smoothing operators (see below) in the Nash inverse function argument can be replaced by Cauchy estimates. Nash's proof of the Ck- case was later extrapolated into the h-principle and Nash–Moser implicit function theorem. A simplified proof of the second Nash embedding theorem was obtained by Günther (1989) who reduced the set of nonlinear partial differential equations to an elliptic system, to which the contraction mapping theorem could be applied. Nash–Kuiper theorem ( C1 embedding theorem) Theorem. Let ( M,g) be a Riemannian manifold and ƒ:Mm → Rn a short C∞-embedding (or immersion) into Euclidean space Rn, where n ≥ m+1. Then for arbitrary ε > 0 there is an embedding (or immersion) m n ƒε : M → R which is (i) in class C1, (ii) isometric: for any two vectors v,w ∈ Tx(M) in the tangent space at x ∈ M, , (iii) ε-close to ƒ: |ƒ(x) − ƒε(x)| < ε for all x ∈ M. In particular, as follows from the Whitney embedding theorem, any m-dimensional Riemannian manifold admits an isometric C1-embedding into an arbitrarily small neighborhood in 2 m-dimensional Euclidean space. The theorem was originally proved by John Nash with the condition n ≥ m+2 instead of n ≥ m+1 and generalized by Nicolaas Kuiper, by a relatively easy trick. The theorem has many counterintuitive implications. For example, it follows that any closed oriented Riemannian surface can be C1 isometrically embedded into an arbitrarily small ball in Euclidean 3-space (from the Gauss–Bonnet theorem, there is no such C2-embedding). And, there exist C 1 isometric embeddings of the hyperbolic plane in R3. http://en.wikipedia.org/wiki/Nash_embedding_theorem 5/23/2011 Nash embedding theorem - Wikipedia, the free encyclopedia Page 2 of 3 Ck embedding theorem The technical statement is as follows: if M is a given m-dimensional Riemannian manifold (analytic or of class C k, 3 ≤ k ≤ ∞), then there exists a number n (with n ≤ m(3m+11)/2 if M is a compact manifold, or n ≤ m(m+1)(3 m+11)/2 if M is a non-compact manifold) and an injective map f : M → Rn (also k analytic or of class C ) such that for every point p of M, the derivative d fp is a linear map from the n tangent space T pM to R which is compatible with the given inner product on T pM and the standard dot product of Rn in the following sense: u, v = dfp(u) · d fp(v) for all vectors u, v in T pM. This is an undetermined system of partial differential equations (PDEs). The Nash embedding theorem is a global theorem in the sense that the whole manifold is embedded into Rn. A local embedding theorem is much simpler and can be proved using the implicit function theorem of advanced calculus. The proof of the global embedding theorem relies on Nash's far-reaching generalization of the implicit function theorem, the Nash–Moser theorem and Newton's method with postconditioning (see ref.). The basic idea of Nash's solution of the embedding problem is the use of Newton's method to prove the existence of a solution to the above system of PDEs. The standard Newton's method fails to converge when applied to the system; Nash uses smoothing operators defined by convolution to make the Newton iteration converge: this is Newton's method with postconditioning. The fact that this technique furnishes a solution is in itself an existence theorem and of independent interest. There is also an older method called Kantorovich iteration that uses Newton's method directly (without the introduction of smoothing operators). References Greene, Robert E.; Jacobowitz, Howard (1971), "Analytic Isometric Embeddings", Annals of Mathematics (The Annals of Mathematics, Vol. 93, No. 1) 93 (1): 189–204, doi:10.2307/1970760, http://jstor.org/stable/1970760 Günther, Matthias (1989), "Zum Einbettungssatz von J. Nash [On the embedding theorem of J. Nash]", Math. Nachr. 144 : 165–187, doi:10.1002/mana.19891440113 Han, Qing; Hong, Jia-Xing (2006), Isometric Embedding of Riemannian Manifolds in Euclidean Spaces , American Mathematical Society, ISBN 0821840711 1 Kuiper, N.H. (1955), "On C -isometric imbeddings I", Nederl. Akad. Wetensch. Proc. Ser. A. 58 : 545–556. 1 Nash, John (1954), "C -isometric imbeddings", Annals of Mathematics (The Annals of Mathematics, Vol. 60, No. 3) 60 (3): 383–396, doi:10.2307/1969840, http://www.jstor.org/stable/1969840. Nash, John (1956), "The imbedding problem for Riemannian manifolds", Annals of Mathematics 63 (1): 20–63, doi:10.2307/1969989, MR0075639, http://www.jstor.org/stable/1969989. Nash, John (1966), "Analyticity of the solutions of implicit function problem with analytic data", Annals of Mathematics (The Annals of Mathematics, Vol. 84, No. 3) 84 (3): 345–355, doi:10.2307/1970448, http://jstor.org/stable/1970448. Retrieved from "http://en.wikipedia.org/wiki/Nash_embedding_theorem" Categories: Riemannian geometry | Mathematical theorems http://en.wikipedia.org/wiki/Nash_embedding_theorem 5/23/2011 Nash embedding theorem - Wikipedia, the free encyclopedia Page 3 of 3 This page was last modified on 17 March 2011 at 19:16. Text is available under the Creative Commons Attribution-ShareAlike License; additional terms may apply. See Terms of Use for details. Wikipedia® is a registered trademark of the Wikimedia Foundation, Inc., a non-profit organization. http://en.wikipedia.org/wiki/Nash_embedding_theorem 5/23/2011.
Load more

Recommended publications
  • Riemannian Submanifolds: a Survey
    RIEMANNIAN SUBMANIFOLDS: A SURVEY BANG-YEN CHEN Contents Chapter 1. Introduction .............................. ...................6 Chapter 2. Nash’s embedding theorem and some related results .........9 2.1. Cartan-Janet’s theorem .......................... ...............10 2.2. Nash’s embedding theorem ......................... .............11 2.3. Isometric immersions with the smallest possible codimension . 8 2.4. Isometric immersions with prescribed Gaussian or Gauss-Kronecker curvature .......................................... ..................12 2.5. Isometric immersions with prescribed mean curvature. ...........13 Chapter 3. Fundamental theorems, basic notions and results ...........14 3.1. Fundamental equations ........................... ..............14 3.2. Fundamental theorems ............................ ..............15 3.3. Basic notions ................................... ................16 3.4. A general inequality ............................. ...............17 3.5. Product immersions .............................. .............. 19 3.6. A relationship between k-Ricci tensor and shape operator . 20 3.7. Completeness of curvature surfaces . ..............22 Chapter 4. Rigidity and reduction theorems . ..............24 4.1. Rigidity ....................................... .................24 4.2. A reduction theorem .............................. ..............25 Chapter 5. Minimal submanifolds ....................... ...............26 arXiv:1307.1875v1 [math.DG] 7 Jul 2013 5.1. First and second variational formulas
    [Show full text]
  • What Can We Do with Nash's Embedding Theorem ?
    SOOCHOW JOURNAL OF MATHEMATICS Volume 30, No. 3, pp. 303-338, July 2004 WHATCANWEDOWITHNASH’SEMBEDDING THEOREM ? BY BANG-YEN CHEN Abstract. According to the celebrated embedding theorem of J. F. Nash, every Riemannian manifold can be isometrically embedded in some Euclidean spaces with sufficiently high codimension. An immediate problem concerning Nash’s theorem is the following: Problem: What can we do with Nash’s embedding theorem ? In other words, what can we do with arbitrary Euclidean submanifolds of arbitrary high codimension if no local or global assumption were imposed on the submanifold ? In this survey, we present some general optimal solutions to this and related prob- lems. We will also present many applications of the solutions to the theory of submanifolds as well as to Riemannian geometry. 1. What Can We Do with Nash’s Embedding Theorem ? According to the celebrated embedding theorem of J. F. Nash [65] published in 1956, every Riemannian manifold can be isometrically embedded in some Eu- clidean spaces with sufficiently high codimension. Received May 27, 2004. AMS Subject Classification. Primary 53C40, 53C42, 53B25; secondary 53A07, 53B21, 53C25, 53D12. Key words. Nash’s embedding theorem, δ-invariant, σ-invariant, warped product manifold, conformally flat submanifold, inequality, real space form, minimal immersion, Lagrangian im- mersion, Einstein manifold. This survey has been delivered at the International Workshop on “Theory of Submanifolds” held at Valenciennes, France from June 25th to 26th and at Leuven, Belgium from June 27th to 28th, 2003. The author would like to express his hearty thanks to Professors F. Dillen, L. Vrancken and L. Verstraelen for organizing this four days conference dedicated to his sixtieth birthday.
    [Show full text]
  • Fundamental Theorems in Mathematics
    SOME FUNDAMENTAL THEOREMS IN MATHEMATICS OLIVER KNILL Abstract. An expository hitchhikers guide to some theorems in mathematics. Criteria for the current list of 243 theorems are whether the result can be formulated elegantly, whether it is beautiful or useful and whether it could serve as a guide [6] without leading to panic. The order is not a ranking but ordered along a time-line when things were writ- ten down. Since [556] stated “a mathematical theorem only becomes beautiful if presented as a crown jewel within a context" we try sometimes to give some context. Of course, any such list of theorems is a matter of personal preferences, taste and limitations. The num- ber of theorems is arbitrary, the initial obvious goal was 42 but that number got eventually surpassed as it is hard to stop, once started. As a compensation, there are 42 “tweetable" theorems with included proofs. More comments on the choice of the theorems is included in an epilogue. For literature on general mathematics, see [193, 189, 29, 235, 254, 619, 412, 138], for history [217, 625, 376, 73, 46, 208, 379, 365, 690, 113, 618, 79, 259, 341], for popular, beautiful or elegant things [12, 529, 201, 182, 17, 672, 673, 44, 204, 190, 245, 446, 616, 303, 201, 2, 127, 146, 128, 502, 261, 172]. For comprehensive overviews in large parts of math- ematics, [74, 165, 166, 51, 593] or predictions on developments [47]. For reflections about mathematics in general [145, 455, 45, 306, 439, 99, 561]. Encyclopedic source examples are [188, 705, 670, 102, 192, 152, 221, 191, 111, 635].
    [Show full text]
  • Embeddings of Statistical Manifolds
    Embeddings of statistical manifolds H^ongV^anL^e Institute of Mathematics, CAS Conference in honor of Shun-ichi Amari Liblice, June 2016 1. Statistical manifold and embedding of statistical manifolds. 2. Obstructions to the existence of an isostatistical immersion. 3. Outline the proof of the existence of an isostatistical embedding. 4. Final remarks and related open problems. 1. Statistical manifold and embedding of statistical manifolds Definition. (Lauritzen, 1987) A statistical manifold( M; g; T ) is a manifold M equipped with a Riemannian metric g and a 3-symmetric tensor T . • We assume here that dim M < 1. Examples. • Statistical model( M; g; T) @ log p@ log p (1) g(ξ; V1;V2) = Ep(·;ξ)( ); @V1 @V2 (2) @ log p @ log p @ log p T(ξ; V1;V1;V1) = Ep(·;ξ)( ): @V1 @V2 @V3 • Manifolds (M;ρ ) where ρ 2 C1(M × M) is a divergence (contrast function). (Eguchi) (M; ρ) =) (g; r; r∗), a torsion-free dualistic structure. Remarks 1. (g; r; r∗) () (g; T ): ∗ T (A; B; C) := g(rAB − rAB; C): 2. Why (M; g; T )? : T is \simpler" than r. Lauritzen's question: some Riemannian manifolds with a symmetric 3-tensor T might not correspond to a particular statistical model. If there exist (Ω; µ) and p :Ω × M ! R conditions hold, we shall call the function p(x; ξ) a probability density for g and T . =) Lauritzen question () the existence question of a probability density for the tensors g and T on a statistical manifold (M; g; T ). Lauritzen's question leads to the immersion problem of statistical manifolds.
    [Show full text]
  • OVERVIEW This Class Is About the H-Principle. This Topic Might Not Have the Huge Name Recogni
    THE H-PRINCIPLE, LECTURES 1 AND 2: OVERVIEW J. FRANCIS This class is about the h-principle. This topic might not have the huge name recognition for you all, which might lead you might think that the h-principle an esoteric or niche topic in mathematics, but it's not. It's a central and useful tool in geometry, and part of the impetus to teach this course is that it's easy to imagine you all making use of the h-principle in your own work. 1. Immersion theory Given two smooth manifolds M and N, an interesting problem is to study smooth maps of M to N. A particularly interesting type of map (if dim(N) ≥ dim(M)) is an embedding, i.e., a smooth math M ! N which is also an injection at the level of sets. Emb(M; N) is a very interesting thing to try to understand; for instance, the study of Emb(S1; R3) is knot thory. Studying embeddings is difficult in part because the property of a map f being an embedding is not local on M. There is a closely related type of map that is local on M, which is an immersion. Recall that f : M ! N is an immersion if for every point x 2 M, the derivative map on tangent spaces dfx : TxM ! Tf(x)N is injective. (Equivalently, f is an immersion if it is locally an embedding, i.e., every point x has an open neighborhood U such that fjU is an embedding.) In the case where the source M is the standard m-disk Dm, there is a fairly simple model for both the spaces of embeddings and immersions into a target N.
    [Show full text]
  • Nash Embedding: a Road Map to Realizing Quantum Hardware 3 Length So As to Generate the Unit Sphere S3 in R4(= C2), and Then Identifying the Antipodal Points on S3
    Nash embedding: a road map to realizing quantum hardware Faisal Shah Khan Quantum Computing, Inc., Leesburg, VA 20175 USA, Department of Mathematics, Khalifa University, Abu Dhabi UAE Email: [email protected] June 18, 2019 Abstract. The non-Euclidean nature of the mathematical model of quantum circuits leaves open the question of their practical implemen- tation in hardware platforms which necessarily reside in the Euclidean space R3. On the other hand, reversible circuits are elements of Euclidean spaces, making their physical realization in hardware platforms possible and practical. Here, the quantum circuit model for quantum computing is mapped into that of reversible computing in a mathematically robust fashion using Nash embedding so that every quantum computation can be realized as an equivalent reversible one. 1 Introduction Now that the industry has started producing and making available software so- lutions that will run on (or will be compatible with) hardware implementing quantum computations, the time is right to start asking whether mathemati- cally formal and robust road-maps to engineering these hardware exist. While I will answer this question in the affirmative in this letter by proposing Nash embedding as one such road-map, I will also raise questions relating to poten- tial applications of Nash embedding that intuitively appear to have a positive answer. Prototype quantum computing systems are currently available from several vendors, including D-Wave Systems and Rigetti Computing (and IBM, for that matter), yet there is still much controversy about whether these qualify as “quan- tum” computers and how accurately they presage eventual quantum computers that will deliver quantum supremacy [1].
    [Show full text]
  • The Embedding Theorems of Whitney and Nash
    GENERAL ARTICLE The Embedding Theorems of Whitney and Nash Harish Seshadri and Kaushal Verma We begin by briefly motivating the idea of a manifold and then discuss the embedding the- orems of Whitney and Nash that allow us to view these objects inside appropriately large Eu- clidean spaces. 1. Introduction (left) Harish Seshadri is with the Department of Let us begin by motivating the concept of a manifold. Mathematics, Indian Start with the unit circle C in the plane given by Institute of x2 + y2 =1. Science, Bengaluru. His research interest is in This is not a graph of a single function y = f(x). How- differential geometry. ever, apart√ from (±1, 0), C is the union of the graphs y ± − x2 (right) Kaushal Verma is of = 1 . Each graph is in one-to-one (bijec- with the Department of tive) correspondence with its domain, namely the inter- Mathematics, Indian val (−1, 1) on the x-axis – apart from the end points Institute of ±1. To take care of points on C near (±1, 0), we see Science, Bengaluru. His that apart from (0, ±1), C is the union of the graphs of research interest is in complex analysis. x = ± 1 − y2. Again, each graph is in bijective cor- respondence with its domain, namely the interval from (0, −1) to (0, 1) on the y-axis. Thus, to accommodate the two exceptional points (±1, 0) we had to change our coordinate axes and hence the functions that identify the pieces of C. Doing all this allows us to describe all points on C in a bijective manner by associating them with points either on the x-axis or the y-axis.
    [Show full text]
  • Nearly Isometric Embedding by Relaxation
    Nearly Isometric Embedding by Relaxation James McQueen Marina Meila˘ Dominique Perrault-Joncas Department of Statistics Department of Statistics Google University of Washington University of Washington Seattle, WA 98103 Seattle, WA 98195 Seattle, WA 98195 [email protected] [email protected] [email protected] Abstract Many manifold learning algorithms aim to create embeddings with low or no dis- tortion (isometric). If the data has intrinsic dimension d, it is often impossible to obtain an isometric embedding in d dimensions, but possible in s>d dimensions. Yet, most geometry preserving algorithms cannot do the latter. This paper pro- poses an embedding algorithm to overcome this. The algorithm accepts as input, besides the dimension d, an embedding dimension s ≥ d. For any data embedding Y, we compute a Loss(Y), based on the push-forward Riemannian metric associ- ated with Y, which measures deviation of Y from from isometry. Riemannian Relaxation iteratively updates Y in order to decrease Loss(Y). The experiments confirm the superiority of our algorithm in obtaining low distortion embeddings. 1 Introduction, background and problem formulation Suppose we observe data points sampled from a smooth manifold M with intrinsic dimension d which is itself a submanifold of D-dimensional Euclidean space M ⊂ RD. The task of manifold learning is to provide a mapping φ : M →N (where N ⊂ Rs) of the manifold into lower dimensional space s ≪ D. According to the Whitney Embedding Theorem [11] we know that M can be embedded smoothly into R2d using one homeomorphism φ. Hence we seek one smooth map φ : M → Rs with d ≤ s ≤ 2d ≪ D.
    [Show full text]
  • Math 535B Final Project 2020: Lecture and Paper
    Math 535b Final Project 2020: Lecture and Paper 1. Overview and timeline The goal of the Math 535b final project is to explore some advanced aspect of Riemannian, symplectic, and/or K¨ahlergeometry (most likely chosen from a list below, but you are welcome to select and plan out a different topic, with instructor approval). You will then give two forms of exposition of this topic: • A 30 minute lecture, delivered to the class. • A 5+ page paper exposition of the topic. Your paper must be typeset (and I strongly recommend LATEX) The timeline for these assignments is: • Your lectures will be scheduled during either the last week of class April 27-May 1 or May 4-8, the week right after. We will schedule precise times in the next two weeks. If you have any scheduling constraints, please let me know ASAP. • Your final assignment is due the last day of class, May 1. If it is necessary, this can be extended somewhat, but please e-mail me in advance to schedule a final (hard) deadline. Some general requirements and/or suggestions: (1) Exposition is important, as are key definitions, intuition, and some key proofs or proof sketches. It is not necessary to give rigorous arguments of every result. (2) The lecture and paper probably should not be identical in emphasis, as the formats are very different. In particular, your goal for the lecture is to convey the relevant ideas at the level your fellow students will understand in lecture. Your target audi- ence for the paper is again your fellow students, but the aim is for your document to be more like a section of a chapter or an expository introductory article.
    [Show full text]
  • The Cartan-Janet Theorem: Local Isometric Embedding of Real-Analytic Riemannian 2-Manifolds Into R3
    The Cartan-Janet Theorem: Local Isometric Embedding of Real-Analytic Riemannian 2-Manifolds into R3 Jesse Madnick August 2, 2014 1 Outline of Notes Introduction to Problem EDS Basics Strategy: Phase 1: Define an EDS I0. Prop 1: Integral manifolds correspond to isometric embeddings Phase 2: Define a larger EDS I. Prop 2: Integral manifolds of I are the same as for I0 Phase 3: Algebraic properties of I Phase 4: Using Cartan-K¨ahlerto show that there exist integral manifolds of I Introduction The goal of these notes is to prove: n 1 Cartan-Janet (1927): Let (M ; g) be a real-analytic Riemannian manifold, N = 2 n(n + 1). N Every point of M has a neighborhood which has a real-analytic isometric embedding into R . We will prove the Cartan-Janet Theorem in the case n = 2, N = 3. The general case is almost exactly analogous to this one, but is much more notationally cumbersome, and requires an addi- tional algebraic lemma. To place this theorem in a broader context, we compare and contrast it with the better-known Nash Embedding Theorem, a global result: Nash Embedding (1956): Let (M n; g) be a Ck Riemannian manifold, where 3 ≤ k ≤ 1 or k = !. k N 1 Then there exists a global C isometric embedding of M into some R , where N ≤ 2 n(3n + 11) if 1 2 M compact, and N ≤ 2 n(3n + 7n + 11) + (2n + 1) if M is non-compact. 1 n N 1 Note: As far as I know, local C embedding of M into R with N = 2 n(n + 1) is open! Remark on the Isometric Embedding PDE n N Let u: M ! R .
    [Show full text]
  • Introduction to Immersion, Embedding, and the Whitney Embedding Theorems
    Introduction to Immersion, Embedding, and the Whitney Embedding Theorems Paul Rapoport November 23, 2015 Abstract We give an overview of immersion in order to present the idea of embedding, then discuss Whit- ney's work on his weak immersion and strong embedding theorems as the main theorems of the talk, drawing briefly on notions of transversality as presented by the previous speaker, rigorously proving a weaker form of Whitney's embedding theorem, and sketching out a result of a linear bound for the minimum dimension of the real space in which a manifold of a given dimension can be embedded. We touch briefly on knot-theoretic applications of embeddings, paying particular attention to alternative arguments for the isotopy of all embeddings of S1 into R4, before finally addressing a recent preliminary examination problem involving embeddings, surveying recent work in the field, and touching on the importance of embedding to two interesting open problems. Introduction In this survey, we will be discussing the concept of immersion as a lead-in to that of embedding, and then finally prove the Whitney embedding theorem for compact manifolds. The Whitney embed- ding theorem says that any smooth, differentiable real manifold with dimension m can be embedded in Rn. The concepts of immersion and embedding, while natural, were only recently fully formal- ized, with Whitney's work finishing as late as 1944. Nevertheless, the concept of an embedding in particular formalizes the intuitive notion of an inclusion map from one smoothly differentiable manifold into another. For simplicity's sake, domains will generally be compact, smoothly differentiable real manifolds, and the codomains of all maps will be some Rn, since we are concerned primarily with classical embedding theory, rather than hyperbolic, complex, or Riemannian embeddings, all of which are mostly beyond the scope of this survey.
    [Show full text]
  • Towards an Algorithmic Realization of Nash's Embedding Theorem
    Towards an Algorithmic Realization of Nash’s Embedding Theorem Nakul Verma CSE, UC San Diego [email protected] Abstract It is well known from differential geometry that an n-dimensional Riemannian manifold can be iso- metrically embedded in a Euclidean space of dimension 2n+1 [Nas54]. Though the proof by Nash is intuitive, it is not clear whether such a construction is achievable by an algorithm that only has access to a finite-size sample from the manifold. In this paper, we study Nash’s construction and develop two algorithms for embedding a fairly general class of n-dimensional Riemannian mani- folds (initially residing in RD) into Rk (where k only depends on some key manifold properties, such as its intrinsic dimension, its volume, and its curvature) that approximately preserves geodesic distances between all pairs of points. The first algorithm we propose is computationally fast and embeds the given manifold approximately isometrically into about O(2cn) dimensions (where c is an absolute constant). The second algorithm, although computationally more involved, attempts to minimize the dimension of the target space and (approximately isometrically) embeds the manifold in about O(n) dimensions. 1 Introduction Finding low-dimensional representations of manifolds has proven to be an important task in data analysis and data visualization. Typically, one wants a low-dimensional embedding to reduce computational costs while maintaining relevant information in the data. For many learning tasks, distances between data-points serve as an important approximation to gauge similarity between the observations. Thus, it comes as no surprise that distance-preserving or isometric embeddings are popular.
    [Show full text]