Riemannian Geometry

🌀Riemannian Geometry Unit 2 – Riemannian Metrics & Induced Distances

Riemannian metrics and induced distances form the backbone of Riemannian geometry. These concepts provide a framework for measuring lengths, angles, and curvature on smooth manifolds, generalizing familiar notions from Euclidean geometry to curved spaces. Understanding these ideas is crucial for studying geodesics, curvature, and geometric flows. They have wide-ranging applications in physics, computer vision, and machine learning, where they help model complex systems and analyze high-dimensional data.

Key Concepts and Definitions

  • Riemannian manifold consists of a smooth manifold MM equipped with a Riemannian metric gg
  • Riemannian metric gg assigns an inner product gpg_p to each tangent space TpMT_pM at every point pMp \in M
    • Inner product varies smoothly as the point pp varies
  • Tangent space TpMT_pM vector space attached to each point pp of the manifold MM
    • Tangent vectors represent velocities of curves passing through the point pp
  • Smooth manifold topological space that locally resembles Euclidean space near each point
    • Transition maps between overlapping coordinate charts are smooth functions
  • Induced distance function d(p,q)d(p,q) measures the shortest path length between points pp and qq on the manifold
  • Geodesic shortest path between two points on the manifold, generalizing the concept of a straight line in Euclidean space
  • Curvature measures the deviation of the manifold from being flat (Euclidean)
    • Quantifies how parallel transport of vectors along closed loops fails to preserve their angle

Riemannian Manifolds Basics

  • Riemannian manifold (M,g)(M, g) fundamental object of study in Riemannian geometry
    • Manifold MM represents the underlying space
    • Riemannian metric gg provides geometric structure
  • Dimension nn of a Riemannian manifold number of coordinates needed to specify a point locally
    • Examples: surface (n=2n=2), space (n=3n=3), spacetime (n=4n=4)
  • Coordinate chart (U,φ)(U, \varphi) consists of an open subset UMU \subset M and a homeomorphism φ:URn\varphi: U \to \mathbb{R}^n
    • Allows expressing points in UU using nn real coordinates
  • Atlas collection of coordinate charts that cover the entire manifold MM
    • Transition maps between overlapping charts must be smooth
  • Tangent bundle TMTM disjoint union of all tangent spaces TpMT_pM of the manifold MM
  • Riemannian manifolds equipped with a natural volume form dVgdV_g induced by the Riemannian metric gg
    • Enables integration of functions over the manifold

Riemannian Metrics: Properties and Examples

  • Riemannian metric gg assigns an inner product ,p\langle \cdot, \cdot \rangle_p to each tangent space TpMT_pM
    • Inner product symmetric, bilinear, and positive-definite
  • Components of the Riemannian metric gijg_{ij} in local coordinates express the inner product of basis vectors i\partial_i and j\partial_j
    • Matrix representation [gij][g_{ij}] symmetric and positive-definite at each point
  • Euclidean metric gEg_E on Rn\mathbb{R}^n assigns the standard dot product to each tangent space
    • Components gij=δijg_{ij} = \delta_{ij} (Kronecker delta)
  • Induced metric on a submanifold SMS \subset M obtained by restricting the ambient Riemannian metric to tangent spaces of SS
    • Example: sphere S2S^2 with induced metric from Euclidean R3\mathbb{R}^3
  • Conformal change of metric multiplying a Riemannian metric gg by a positive smooth function ff
    • Preserves angles between vectors but changes lengths
  • Product metric on a product manifold M×NM \times N obtained by summing the metrics on MM and NN
    • Example: torus S1×S1S^1 \times S^1 with product metric from circle S1S^1
  • Riemannian submersion π:(M,gM)(N,gN)\pi: (M, g_M) \to (N, g_N) maps Riemannian manifolds such that the differential dπd\pi is an isometry on horizontal subspaces
    • Useful for constructing new Riemannian manifolds with symmetries

Induced Distances and Length Spaces

  • Riemannian metric gg induces a distance function d(p,q)d(p,q) on the manifold MM
    • Measures the infimum of lengths of piecewise smooth curves connecting pp and qq
  • Length of a piecewise smooth curve γ:[a,b]M\gamma: [a,b] \to M given by the integral of the norm of its velocity vector γ(t)\gamma'(t)
    • L(γ)=abgγ(t)(γ(t),γ(t))dtL(\gamma) = \int_a^b \sqrt{g_{\gamma(t)}(\gamma'(t), \gamma'(t))} dt
  • Riemannian manifold (M,g)(M, g) becomes a metric space (M,d)(M, d) with the induced distance function dd
    • Satisfies axioms of a metric: non-negativity, symmetry, and triangle inequality
  • Hopf-Rinow theorem states that a connected Riemannian manifold is complete as a metric space if and only if it is geodesically complete
    • Completeness ensures the existence of minimal geodesics between any two points
  • Length space (X,d)(X, d) generalizes the concept of a Riemannian manifold
    • Distance function dd defined as the infimum of lengths of admissible paths
  • Intrinsic metric on a length space induced by the length structure
    • Makes the space a metric space and generalizes the Riemannian distance
  • Alexandrov space length space with curvature bounded below in the sense of comparison triangles
    • Generalizes Riemannian manifolds with lower curvature bounds

Geodesics and Shortest Paths

  • Geodesic γ:IM\gamma: I \to M locally length-minimizing curve on a Riemannian manifold (M,g)(M, g)
    • Satisfies the geodesic equation γγ=0\nabla_{\gamma'}\gamma' = 0, where \nabla is the Levi-Civita connection
  • Exponential map expp:TpMM\exp_p: T_pM \to M maps tangent vectors at pp to points on the manifold along geodesics
    • Geodesics through pp correspond to straight lines in TpMT_pM under the exponential map
  • Minimizing geodesic shortest path between two points on the manifold
    • May not be unique, as in the case of antipodal points on a sphere
  • Cut locus CpC_p of a point pp set of points where geodesics from pp cease to be minimizing
    • Exponential map is a diffeomorphism from a star-shaped domain in TpMT_pM to MCpM \setminus C_p
  • Geodesic triangle formed by three minimizing geodesics connecting three points on the manifold
    • Angles and side lengths satisfy triangle comparison theorems based on curvature bounds
  • Jacobi fields measure the deviation of nearby geodesics
    • Satisfy the Jacobi equation, a second-order linear ODE involving the Riemann curvature tensor
  • Rauch comparison theorem relates the growth of Jacobi fields to the curvature of the manifold
    • Used to prove the Toponogov triangle comparison theorem and the sphere theorem

Curvature in Riemannian Geometry

  • Riemann curvature tensor R(X,Y)ZR(X, Y)Z measures the non-commutativity of the covariant derivative \nabla
    • Quantifies the failure of parallel transport along infinitesimal parallelograms
  • Sectional curvature K(X,Y)K(X, Y) measures the Gaussian curvature of the 2-plane spanned by orthonormal vectors XX and YY
    • Computed as K(X,Y)=R(X,Y)Y,XX,XY,YX,Y2K(X, Y) = \frac{\langle R(X, Y)Y, X \rangle}{\langle X, X \rangle \langle Y, Y \rangle - \langle X, Y \rangle^2}
  • Ricci curvature Ric(X,X)\text{Ric}(X, X) trace of the Riemann curvature tensor in the direction of the vector XX
    • Measures the average sectional curvature of planes containing XX
  • Scalar curvature SS contraction of the Ricci curvature with the metric tensor
    • Provides a single real number measuring the overall curvature at each point
  • Constant curvature manifolds have the same sectional curvature KK at every point and for every 2-plane
    • Examples: Euclidean space (K=0K=0), sphere (K>0K>0), hyperbolic space (K<0K<0)
  • Comparison theorems relate the geometry of a Riemannian manifold to that of constant curvature spaces
    • Bonnet-Myers theorem, Synge's theorem, and the Cartan-Hadamard theorem
  • Curvature affects the behavior of geodesics, Jacobi fields, and the topology of the manifold
    • Positive curvature implies compact manifolds and converging geodesics
    • Negative curvature implies non-compact manifolds and diverging geodesics

Applications and Real-World Examples

  • General relativity models spacetime as a 4-dimensional Lorentzian manifold with a metric tensor
    • Curvature of spacetime related to the presence of matter and energy via Einstein's field equations
  • Riemannian geometry used in computer vision and image processing
    • Treats images as 2D manifolds and uses geodesic distances for segmentation and registration
  • Machine learning and data analysis utilize Riemannian geometry
    • Manifold learning algorithms (Isomap, LLE) seek to uncover the underlying low-dimensional structure of high-dimensional data
  • Robotics and motion planning benefit from Riemannian geometry
    • Configuration spaces of robots modeled as Riemannian manifolds
    • Geodesics used to plan optimal trajectories and avoid obstacles
  • Geometric mechanics formulates classical mechanics on Riemannian manifolds
    • Configuration space of a mechanical system treated as a Riemannian manifold with a kinetic energy metric
  • Lie groups equipped with bi-invariant Riemannian metrics
    • Geodesics correspond to one-parameter subgroups and the exponential map is the Lie exponential
  • Optimization on Riemannian manifolds generalizes classical optimization techniques
    • Gradient descent, conjugate gradient, and trust-region methods adapted to curved spaces
  • Shape analysis and computational anatomy study the geometry of shapes and anatomical structures
    • Shapes represented as points on infinite-dimensional Riemannian manifolds of diffeomorphisms or embeddings

Advanced Topics and Further Reading

  • Comparison geometry studies Riemannian manifolds with curvature bounds
    • Alexandrov spaces, CAT(kk) spaces, and Gromov-Hausdorff convergence
  • Geometric flows evolve Riemannian metrics according to PDEs
    • Ricci flow, mean curvature flow, and Yamabe flow
    • Used to prove the Poincaré conjecture and the geometrization conjecture
  • Harmonic maps between Riemannian manifolds generalize the concept of harmonic functions
    • Minimize the energy functional and satisfy the harmonic map equation
  • Eigenvalue problems on Riemannian manifolds study the spectrum of the Laplace-Beltrami operator
    • Relate the geometry of the manifold to the properties of eigenfunctions and eigenvalues
  • Spin geometry studies Riemannian manifolds with additional structure called a spin structure
    • Allows the construction of spinor bundles and Dirac operators
  • Kähler geometry studies complex manifolds with a compatible Riemannian metric
    • Unifies Riemannian, complex, and symplectic geometry
  • Finsler geometry generalizes Riemannian geometry by allowing the metric to depend on direction
    • Finsler metrics assign a norm to each tangent space instead of an inner product
  • Sub-Riemannian geometry studies manifolds with a distribution of subspaces of the tangent spaces
    • Geodesics are constrained to be tangent to the distribution
  • Further reading:
    • "Riemannian Geometry" by Manfredo P. do Carmo
    • "Riemannian Manifolds: An Introduction to Curvature" by John M. Lee
    • "Comparison Geometry" by Karsten Grove
    • "Geometric Flows and Geometric Operators" by Bennett Chow and Dan Knopf


© 2024 Fiveable Inc. All rights reserved.
AP® and SAT® are trademarks registered by the College Board, which is not affiliated with, and does not endorse this website.

© 2024 Fiveable Inc. All rights reserved.
AP® and SAT® are trademarks registered by the College Board, which is not affiliated with, and does not endorse this website.