Source author record

Chi-Wai Leung

Explore this author’s publications, collaborators and research topics. This catalog record has not been claimed by its author.

ResearcherUnclaimed source record

Catalog footprint

Research overview

9works
10topics
2close collaborators

Published work

9 publications

preprint2020arXiv

Analytic bundle structure on the idempotent manifold

Let XX be a (real or complex) Banach space, and I(X)\mathcal{I}(X) be the set of all (non-zero and non-identity) idempotents; i.e., bounded linear operators on XX whose squares equal themselves. We show that the Banach submanifold I(X)\mathcal{I}(X) of L(X)\mathcal{L}(X) is a locally trivial analytic affine-Banach bundle over the Grassmann manifold G(X)\mathscr{G}(X), via the map κκ that sends Q∈I(X)Q\in \mathcal{I}(X) to Q(X)Q(X), such that the affine-Banach space structure on each fiber is the one induced from L(X)\mathcal{L}(X) (in particular, every fiber is an affine-Banach subspace of L(X)\mathcal{L}(X)). Using this, we show that if KK is a real Hilbert space, then the assignment (E,T)↦T∗∘PE⊥+PE, where E∈G(K) and T∈L(E,E⊥),(E,T)\mapsto T^*\circ P_{E^\bot} + P_{E}, \quad \text{ where } E\in \mathscr{G}(K)\text{ and } T\in \mathcal{L}(E,E^\bot), induces a bi-analytic bijection from the total space of the tangent bundle, T(G(K))\mathbf{T}(\mathscr{G}(K)), of G(K)\mathscr{G}(K) onto I(K)\mathcal{I}(K) (here, E⊥E^\bot is the orthogonal complement of EE, PE∈L(K)P_E\in \mathcal{L}(K) is the orthogonal projection onto EE, and T∗T^* is the adjoint of TT). Notice that this bi-analytic bijection is an affine map on each tangent plane.

preprint2020arXiv

On a variant of Tingley's problem for some function spaces

Let (Ω,A,μ)(Ω, \mathfrak{A}, μ) and (Γ,B,ν)(Γ, \mathfrak{B}, ν) be two arbitrary measure spaces, and p∈[1,∞]p\in [1,\infty]. Set Lp(μ)+sp:={f∈Lp(μ):∥f∥p=1;f≥0 μ-a.e.}L^p(μ)_+^\mathrm{sp}:= \{f\in L^p(μ): \|f\|_p =1; f\geq 0\ μ\text{-a.e.} \} i.e., the positive part of the unit sphere of Lp(μ)L^p(μ). We show that every metric preserving bijection Φ:Lp(μ)+sp→Lp(ν)+spΦ: L^p(μ)_+^\mathrm{sp} \to L^p(ν)_+^\mathrm{sp} can be extended (necessarily uniquely) to an isometric order isomorphism from Lp(μ)L^p(μ) onto Lp(ν)L^p(ν). A Lamperti form, i.e., a weighted composition like form, of ΦΦ is provided, when (Γ,B,ν)(Γ, \mathfrak{B}, ν) is localizable (in particular, when it is σσ-finite). On the other hand, we show that for compact Hausdorff spaces XX and YY, if ΦΦ is a metric preserving bijection from the positive part of the unit sphere of C(X)C(X) to that of C(Y)C(Y), then there is a homeomorphism τ:Y→Xτ:Y\to X satisfying Φ(f)(y)=f(τ(y))Φ(f)(y) = f(τ(y)) (f∈C(X)+sp;y∈Yf\in C(X)_+^\mathrm{sp}; y\in Y).

preprint2015arXiv

Transition probabilities of normal states determine the Jordan structure of a quantum system

Let Φ:S(M1)→S(M2)Φ:\mathfrak{S}(M_1)\to \mathfrak{S}(M_2) be a bijection (not assumed affine nor continuous) between the sets of normal states of two quantum systems, modelled on the self-adjoint parts of von Neumann algebras M1M_1 and M2M_2, respectively. This paper concerns with the situation when ΦΦ preserves (or partially preserves) one of the following three notions of "transition probability" on the normal state spaces: the Uhlmann transition probability PUP_U, the Raggio transition probability PBP_B and an "asymmetric transition probability" P0P_0 as defined in this article. It is shown that the two systems are isomorphic, i.e. M1M_1 and M2M_2 are Jordan ∗^*-isomorphic, if ΦΦ preserves all pairs with zero Uhlmann (respectively, Raggio or asymmetric) transition probability, i.e., for any normal states μμ and νν, we have P(Φ(μ),Φ(ν))=0if and only ifP(μ,ν)=0, P\big(Φ(μ),Φ(ν)\big) = 0 \quad \text{if and only if} \quad P(μ,ν)=0, where PP stands for PUP_U (respectively, PRP_R or P0P_0). Furthermore, as an extension of Wigner's theorem, it is shown that there is a Jordan ∗^*-isomorphism Θ:M2→M1Θ:M_2\to M_1 with Φ=Θ∗∣S(M1)Φ= Θ^*|_{\mathfrak{S}(M_1)} if and only if ΦΦ preserves the "asymmetric transition probability". This is also equivalent to ΦΦ preserving the Raggio transition probability. Consequently, if ΦΦ preserves the Raggio transition probability, it will preserve the Uhlmann transition probability as well. As another application, the sets of normal states equipped with either the usual metric, the Bures metric or "the metric induced by the self-dual cone" are complete Jordan ∗^*-invariants for the underlying von Neumann algebras.

preprint2014arXiv

Berkovich spectra of elements in Banach Rings

Adapting the notion of the spectrum ΣaΣ_a for an element aa in an ultrametric Banach algebra (as defined by Berkovich), we introduce and briefly study the Berkovich spectrum σRBer(u)σ^{Ber}_R(u) of an element uu in a Banach ring RR. This spectrum is a compact subset of the affine analytic space AZ1A_Z^1 over ZZ, and the later can be identified with the "equivalence classes" of all elements in all complete valuation fields. If RR is generated by uu as a unital Banach ring, then σRBer(u)σ^{Ber}_R(u) coincides with the spectrum of RR (as defined by Berkovich). If RR is a unital complex Banach algebra, then σRBer(u)σ^{Ber}_R(u) is the "folding up" of the usual spectrum σB(u)σ_B(u) alone the real axis. For a non-Archimedean complete valuation field kk and an infinite dimensional ultrametric kk-Banach space EE with an orthogonal base, if u∈L(E)u\in L(E) is a completely continuous operator, we show that many different ways to define the spectrum of uu give the same compact set σL(E)Ber(u)σ^{Ber}_{L(E)}(u). As an application, we give a lower bound for the valuations of the zeros of the Fredholm determinant det⁡(1−t⋅u)\det(1- t\cdot u) (as defined by Serre) in complete valuation field extensions of kk. Using this, we give a concrete example of a completely continuous operator whose Fredholm determinant does not have any zero in any complete valuation field extension of kk.

preprint2010arXiv

Automatic continuity and C0(Ω)C_0(Ω)-linearity of linear maps between C0(Ω)C_0(Ω)-modules

Let ΩΩ be a locally compact Hausdorff space. We show that any local C\mathbb{C}-linear map (where "local" is a weaker notion than C0(Ω)C_0(Ω)-linearity) between Banach C0(Ω)C_0(Ω)-modules are "nearly C0(Ω)C_0(Ω)-linear" and "nearly bounded". As an application, a local C\mathbb{C}-linear map θθ between Hilbert C0(Ω)C_0(Ω)-modules is automatically C0(Ω)C_0(Ω)-linear. If, in addition, ΩΩ contains no isolated point, then any C0(Ω)C_0(Ω)-linear map between Hilbert C0(Ω)C_0(Ω)-modules is automatically bounded. Another application is that if a sequence of maps {θn}\{θ_n\} between two Banach spaces "preserve c0c_0-sequences" (or "preserve ultra-c0c_0-sequences"), then θnθ_n is bounded for large enough nn and they have a common bound. Moreover, we will show that if θθ is a bijective "biseparating" linear map from a "full" essential Banach C0(Ω)C_0(Ω)-module EE into a "full" Hilbert C0(Δ)C_0(Δ)-module FF (where ΔΔ is another locally compact Hausdorff space), then θθ is "nearly bounded" (in fact, it is automatically bounded if ΔΔ or ΩΩ contains no isolated point) and there exists a homeomorphism σ:Δ→Ωσ: Δ\rightarrow Ω such that θ(e⋅φ)=θ(e)⋅φ∘σθ(e\cdot φ) = θ(e)\cdot φ\circ σ (e∈E,φ∈C0(Ω)e\in E, φ\in C_0(Ω)).

preprint2010arXiv

Linear orthogonality preservers of Hilbert C∗C^*-modules over general C∗C^*-algebras

As a partial generalisation of the Uhlhorn theorem to Hilbert C∗C^*-modules, we show in this article that the module structure and the orthogonality structure of a Hilbert C∗C^*-module determine its Hilbert C∗C^*-module structure. In fact, we have a more general result as follows. Let AA be a C∗C^*-algebra, EE and FF be Hilbert AA-modules, and IEI_E be the ideal of AA generated by {⟨x,y⟩A:x,y∈E}\{\langle x,y\rangle_A: x,y\in E\}. If Φ:E→FΦ: E\to F is an AA-module map, not assumed to be bounded but satisfying ⟨Φ(x),Φ(y)⟩A = 0whenever⟨x,y⟩A = 0, \langle Φ(x),Φ(y)\rangle_A\ =\ 0\quad\text{whenever}\quad\langle x,y\rangle_A\ =\ 0, then there exists a unique central positive multiplier u∈M(IE)u\in M(I_E) such that ⟨Φ(x),Φ(y)⟩A = u⟨x,y⟩A(x,y∈E). \langle Φ(x), Φ(y)\rangle_A\ =\ u \langle x, y\rangle_A\qquad (x,y\in E). As a consequence, ΦΦ is automatically bounded, the induced map Φ0:E→Φ(E)‾Φ_0: E\to \overline{Φ(E)} is adjointable, and Eu1/2‾\overline{Eu^{1/2}} is isomorphic to Φ(E)‾\overline{Φ(E)} as Hilbert AA-modules. If, in addition, ΦΦ is bijective, then EE is isomorphic to FF.

preprint2010arXiv

Linear orthogonality preservers of Hilbert bundles

Due to the corresponding fact concerning Hilbert spaces, it is natural to ask if the linearity and the orthogonality structure of a Hilbert C∗C^*-module determine its C∗C^*-algebra-valued inner product. We verify this in the case when the C∗C^*-algebra is commutative (or equivalently, we consider a Hilbert bundle over a locally compact Hausdorff space). More precisely, a C\mathbb{C}-linear map θθ (not assumed to be bounded) between two Hilbert C∗C^*-modules is said to be "orthogonality preserving" if <θ(x),θ(y)>=0\left<θ(x),θ(y)\right> =0 whenever <x,y>=0\left<x,y\right> =0. We prove that if θθ is an orthogonality preserving map from a full Hilbert C0(Ω)C_0(Ω)-module EE into another Hilbert C0(Ω)C_0(Ω)-module FF that satisfies a weaker notion of C0(Ω)C_0(Ω)-linearity (known as "localness"), then θθ is bounded and there exists ϕ∈Cb(Ω)+ϕ\in C_b(Ω)_+ such that <θ(x),θ(y)> = ϕ⋅<x,y>,∀x,y∈E. \left<θ(x),θ(y)\right>\ =\ ϕ\cdot\left<x,y\right>, \quad \forall x,y \in E. On the other hand, if FF is a full Hilbert C∗C^*-module over another commutative C∗C^*-algebra C0(Δ)C_0(Δ), we show that a "bi-orthogonality preserving" bijective map θθ with some "local-type property" will be bounded and satisfy <θ(x),θ(y)> = ϕ⋅<x,y>∘σ,∀x,y∈E \left<θ(x),θ(y)\right>\ =\ ϕ\cdot\left<x,y\right>\circσ, \quad \forall x,y \in E where ϕ∈Cb(Ω)+ϕ\in C_b(Ω)_+ and σ:Δ→Ωσ: Δ\rightarrow Ω is a homeomorphism.

Research graph

See the researcher in context

Open full explorer

Inspect adjacent papers, topics, institutions and collaborators without losing the researcher page.

Building this map preview

BZPEER is loading the nearby papers, people, topics and institutions for this page.