Infinite-Dimensional Heisenberg Groups

• Infinite-dimensional Heisenberg groups are step-2 stratified Banach–Lie groups formed as central extensions of infinite-dimensional abelian groups using skew-symmetric bilinear forms.
• They feature canonical sub-Riemannian and weak Riemannian geometries, enabling analysis via CC distances, Gaussian measures, and dimension-free functional inequalities.
• Their representation theory includes infinite-dimensional Weyl–Schrödinger representations and Wigner transforms, underpinning applications in harmonic analysis, stochastic processes, and sub-Riemannian geometry.

Infinite-dimensional Heisenberg groups are step-2 stratified Banach–Lie groups arising as central extensions of infinite-dimensional abelian groups (typically, Banach or Hilbert spaces) by finite- or infinite-dimensional centers via skew-symmetric bilinear forms. These groups generalize the classical, finite-dimensional Heisenberg group and play central roles in infinite-dimensional harmonic analysis, representation theory, stochastic analysis on groups, and sub-Riemannian geometry, particularly in the presence of Gaussian measures. They are equipped with canonical sub-Riemannian and often degenerate weak Riemannian geometries, exhibit dimension-free functional inequalities, and support infinite-dimensional analogues of Weyl–Schrödinger representations, Wigner transforms, and Taylor/Fock isomorphisms.

1. Algebraic and Manifold Structure

Infinite-dimensional Heisenberg groups are typically modeled as $W \times \mathbf{C}$, where $W$ is a real separable Banach space and $\mathbf{C}$ is a finite-dimensional real (or complex) inner-product space. The group law is

$$(w_1, c_1) \cdot (w_2, c_2) = (w_1 + w_2,\, c_1 + c_2 + \tfrac12\, \omega(w_1,w_2))$$

where $\omega: W \times W \to \mathbf{C}$ is a continuous (and often surjective) skew-symmetric bilinear form. The Lie algebra $\mathfrak{g} = W \oplus \mathbf{C}$ has bracket $[(w_1, c_1), (w_2, c_2)] = (0,\, \omega(w_1, w_2))$, yielding a step-2 nilpotent stratification $\mathfrak{g} = V_1 \oplus V_2$ with $V_1 = W$, $V_2 = \mathbf{C}$ (Eldredge et al., 2023).

These groups admit a Banach–Lie structure when the commutator subgroup $W$0 is locally compact, with exponential charts and scalable dilations $W$1. Typical models involve abstract Wiener spaces $W$2, with Gaussian measure $W$3 and dense Cameron–Martin Hilbert subspace $W$4, such that the bracket-generating (Hörmander) condition is met for $W$5 restricted to $W$6 (Driver et al., 2013, Gordina et al., 2011).

2. Sub-Riemannian and Weak Riemannian Geometry

The canonical sub-Riemannian structure is defined by declaring the horizontal distribution $W$7 (or $W$8 when restricted to the Cameron–Martin directions) and equipping it with the Hilbert norm of $W$9. The Carnot–Carathéodory (CC) distance is induced by piecewise horizontal curves. The presence of a nondegenerate bracket condition on $\mathbf{C}$0 ensures the applicability of stochastic and geometric analysis, as in the hypoelliptic framework (Gordina et al., 2021, Baudoin et al., 2011).

Weak Riemannian metrics may be introduced via inner products induced by compact, strictly positive operators $\mathbf{C}$1 on $\mathbf{C}$2 or $\mathbf{C}$3, as in $\mathbf{C}$4 with $\mathbf{C}$5 trace-class, making these inner products weaker than the ambient Banach space norm (Magnani et al., 2021). When extended to the entire group as left-invariant, such metrics induce degenerate geodesic distances, a manifestation of the "Michor–Mumford phenomenon"—the metric topology becomes strictly weaker than the manifold topology, causing the Riemannian and sub-Riemannian distances to vanish on large sets (Magnani et al., 2021).

3. Measures, Heat Kernels, and Functional Inequalities

Canonical Gaussian measures on $\mathbf{C}$6 (the abstract Wiener measure) serve as the background for stochastic and analytic studies. The heat kernel measures $\mathbf{C}$7 on the full group (or on reduced quotients) arise as the law of stochastic processes

$\mathbf{C}$8

where $\mathbf{C}$9 is $$(w_1, c_1) \cdot (w_2, c_2) = (w_1 + w_2,\, c_1 + c_2 + \tfrac12\, \omega(w_1,w_2))$$0-valued Brownian motion with Cameron–Martin covariance (Baudoin et al., 2011, Driver et al., 2013, Dobbs et al., 2012).

These heat kernel measures are strictly positive, quasi-invariant under Cameron–Martin translations, and possess $$(w_1, c_1) \cdot (w_2, c_2) = (w_1 + w_2,\, c_1 + c_2 + \tfrac12\, \omega(w_1,w_2))$$1 bounds on Radon–Nikodym derivatives given explicit in terms of CC distances and curvature–dimension parameters (Baudoin et al., 2011). They are shown to satisfy strong smoothness in the sense of infinite-order integration by parts (Malliavin calculus), both on the group and path space (Dobbs et al., 2012), and the measures are absolutely continuous with respect to the Gaussian–Lebesgue product measure with explicit densities (Driver et al., 2013).

A significant property is the validity of dimension-free logarithmic Sobolev inequalities (LSI): for suitable $$(w_1, c_1) \cdot (w_2, c_2) = (w_1 + w_2,\, c_1 + c_2 + \tfrac12\, \omega(w_1,w_2))$$2 and all $$(w_1, c_1) \cdot (w_2, c_2) = (w_1 + w_2,\, c_1 + c_2 + \tfrac12\, \omega(w_1,w_2))$$3,

$$(w_1, c_1) \cdot (w_2, c_2) = (w_1 + w_2,\, c_1 + c_2 + \tfrac12\, \omega(w_1,w_2))$$4

with universal (sharp) $$(w_1, c_1) \cdot (w_2, c_2) = (w_1 + w_2,\, c_1 + c_2 + \tfrac12\, \omega(w_1,w_2))$$5 matching the finite-dimensional Heisenberg group (Gordina et al., 2021, Gordina et al., 3 Dec 2025). The functional inequalities transfer to reduced Heisenberg groups (central quotients) via quasi-homeomorphism arguments (Gordina et al., 3 Dec 2025).

4. Representation Theory: Weyl–Schrödinger Representations and Wigner Theory

Infinite-dimensional Heisenberg groups admit irreducible Weyl–Schrödinger representations on spaces of square-integrable functions over Gaussian or invariant projective-limit measures. For $$(w_1, c_1) \cdot (w_2, c_2) = (w_1 + w_2,\, c_1 + c_2 + \tfrac12\, \omega(w_1,w_2))$$6 a Hilbert space, the group $$(w_1, c_1) \cdot (w_2, c_2) = (w_1 + w_2,\, c_1 + c_2 + \tfrac12\, \omega(w_1,w_2))$$7 acts on $$(w_1, c_1) \cdot (w_2, c_2) = (w_1 + w_2,\, c_1 + c_2 + \tfrac12\, \omega(w_1,w_2))$$8, where $$(w_1, c_1) \cdot (w_2, c_2) = (w_1 + w_2,\, c_1 + c_2 + \tfrac12\, \omega(w_1,w_2))$$9 is a $\omega: W \times W \to \mathbf{C}$0-invariant Radon measure on the projective limit $\omega: W \times W \to \mathbf{C}$1 of finite-dimensional unitary groups (Lopushansky, 2019, Lopushansky, 2017).

The representation is constructed via shift and multiplicative operators on Hardy or Fock–Hilbert–Schmidt spaces $\omega: W \times W \to \mathbf{C}$2 of entire functions, with the Weyl operator $\omega: W \times W \to \mathbf{C}$3 realizing the canonical commutation relations: $\omega: W \times W \to \mathbf{C}$4 and central elements acting by scalars. These representations are irreducible and satisfy generalized Stone–von Neumann uniqueness (Lopushansky, 2019, Lopushansky, 2017, Beltita et al., 2015).

The infinite-dimensional Wigner transform $\omega: W \times W \to \mathbf{C}$5 is defined on $\omega: W \times W \to \mathbf{C}$6 and is unitary onto its image in $\omega: W \times W \to \mathbf{C}$7 of the infinite-dimensional phase space, satisfying orthogonality and covariance analogous to the finite-dimensional setting (Beltita et al., 2015).

5. Holomorphic Function Theory, Taylor/Fock Isomorphism, and Analytic Structures

Segal–Bargmann-type or Taylor isomorphisms generalize to these groups, establishing unitary equivalence between $\omega: W \times W \to \mathbf{C}$8-holomorphic functions on the group (with respect to heat kernel measure) and completions of the universal enveloping algebra of the Cameron–Martin Lie subalgebra ("non-commutative Fock space") (Gordina et al., 2011). The isomorphism is given as a composition of restriction to the Cameron–Martin subgroup and Taylor expansion at the identity, with norm matching via the Fock-type Hilbert structure. The expansion yields convergent series for functions in the $\omega: W \times W \to \mathbf{C}$9-space and connects the analytic geometry of the group with representation algebra.

In the complexified case, Paley–Wiener isomorphisms relate Hardy/Fock analytic function spaces and $\mathfrak{g} = W \oplus \mathbf{C}$0 spaces over the virtual unitary group, with explicit Schur and power-sum polynomial bases and Fourier–Laplace transforms (Lopushansky, 2017, Lopushansky, 2019).

6. Measure-Theoretic and Geometric Notions: Null Sets and Banach–Lie Manifold Properties

Infinite-dimensional Heisenberg-like groups exhibit Banach–Lie manifold structures when their commutator subgroup is finite-dimensional and locally compact (Eldredge et al., 2023). The geometric and measure-theoretic structure is captured via scalable dilations, complete gauge distances, and closure under Carnot subgroups.

Multiple notions of "null sets" arise: Aronszajn null (directional smallness), CAC-null (with respect to convolutions of absolutely continuous measures on Carnot subgroups), and null for heat kernel measures are all equivalent in the Heisenberg setting (Eldredge et al., 2023). The heat kernel is quasi-invariant under Cameron–Martin translations, and every Aronszajn-null set is also heat kernel null. This interplay is vital in stochastic analysis, geometric measure theory, and the study of differentiability and quasi-invariance in infinite-dimensional Lie groups.

7. Open Problems and Directions

Key open questions include the extension of the equivalence of null set notions to higher step Carnot/Banach–Lie groups, establishing curvature-dimension estimates, constructing holomorphic and analytic invariants beyond the step-2 setting, and explicit characterization of left-invariant $\mathfrak{g} = W \oplus \mathbf{C}$1-ideals of null sets in infinite-dimensional Carnot groups. Further directions concern the behavior of Taylor/Fock constants, extension to more general central extensions, and applications to stochastic PDEs and infinite-dimensional geometric analysis (Dobbs et al., 2012, Eldredge et al., 2023).

---

References

• (Gordina et al., 2011) A subelliptic Taylor isomorphism on infinite-dimensional Heisenberg groups.
• (Baudoin et al., 2011) Quasi-invariance for heat kernel measures on sub-Riemannian infinite-dimensional Heisenberg groups.
• (Dobbs et al., 2012) Smoothness of Heat Kernel Measures on Infinite-Dimensional Heisenberg-Like Groups.
• (Driver et al., 2013) Hypoelliptic heat kernels on infinite-dimensional Heisenberg groups.
• (Beltita et al., 2015) On Wigner transforms in infinite dimensions.
• (Lopushansky, 2017) Paley-Wiener isomorphism over infinite-dimensional unitary groups.
• (Lopushansky, 2019) Weyl-Schrödinger representations of Heisenberg groups in infinite dimensions.
• (Gordina et al., 2021) Logarithmic Sobolev inequalities on non-isotropic Heisenberg groups.
• (Magnani et al., 2021) On the Michor-Mumford phenomenon in the infinite dimensional Heisenberg group.
• (Eldredge et al., 2023) Notions of null sets in infinite-dimensional Carnot groups.
• (Gordina et al., 3 Dec 2025) Logarithmic Sobolev inequalities on infinite-dimensional reduced Heisenberg groups.