home *** CD-ROM | disk | FTP | other *** search
- Xref: sparky sci.physics:22421 alt.sci.physics.new-theories:2721
- Newsgroups: sci.physics,alt.sci.physics.new-theories
- Path: sparky!uunet!well!sarfatti
- From: sarfatti@well.sf.ca.us (Jack Sarfatti)
- Subject: Wavelets, squeezed states 7 Reproducing Kernels
- Message-ID: <C0MC1o.A5E@well.sf.ca.us>
- Sender: news@well.sf.ca.us
- Organization: Whole Earth 'Lectronic Link
- Date: Sun, 10 Jan 1993 03:22:35 GMT
- Lines: 126
-
-
- 7. Reproducing Kernels
- Nothing here is original it's all G Kaiser's stuff.
-
- K(m,m') is called a "reproducing kernel". M is an arbitrary set. A set of
- functions g(m) on M forms a Hilbert space F under some inner product <|>.
- The exact form of the inner product does not matter.
-
- Let K(m,m') be complex-valued on MxM such that
-
- (42) for every m in M, Km(m') = K(m',m) in F.
-
- (43) for every m in M and every g in F
-
- g(m) = <K(m,m)|g> = <Km|g> (p.30)
-
- then F is a reproducing-kernel Hilbert space and K(m,m') is its reproducing
- kernel.
-
- K(m',m) = <K(m',m')|K(m,m)> = <Km'|Km> (44)
-
- K(m,m')* = K(m',m> (45)
-
- K(m,m) = |Km|^2 >= 0 all m (46)
-
- |K(m,m')| <= |Km||Km'| (47) Schwartz inequality
-
- The kernel function K itself virtually generates the whole structure. All
- the functions in F inherit the boundedness and growth properties of K. If K
- has singularities, then some functions in F have those singularities.
-
- *If M is a classical phase space, F is the corresponding Hilbert space of
- the quantum system and |g(m)|^2 is the probability density in the quantum
- state g of finding the system in the classical state m.
-
- This assertion seems rather important!
-
- *Basic idea of coherent and squeezed states:
- The quantum state which maximizes the probability density of being in the
- classical state mo is modulo over all phase gmo. That is, gmo is a quantum
- wave packet which is optimally localized at the classical state (i.e.,
- point) in phase space.
-
-
- *Contact of this stuff to new quantum computer hardware? (half-baked
- Sarfatti speculation or poetic precognition channeled from Q as in Star
- Trek The Next Generation)
-
- Sarfatti problem: How do we use this for fermions? Need anticommuting
- Grassmann variables? What does classical state mean for fermions? Is this
- stuff only good for boson? What about fractional or para-statistics or
- anyons as in 2-d solid state (quantum wells) fractional Hall effect? What
- about quantum dots?
-
- The reproducing kernel embodies the entire Hilbert space. Properties of K
- are imaged in properties of g in F.
-
- Given Hilbert space F whose elements are all functions on M, how do we know
- whether F has a reproducing kernel K? A necessary conditiion is
-
- |f(m)| = |<Km|f>| <= |Km||f| , all f in F and all m in M (48)
-
- For fixed m, there is an "evaluation map" Em of Hilbert space F to complex
- plane C defined by
-
- Em(f) = f(m) (49)
-
- is a bounded linear functional on F.
-
- Riesz representation theorem implies
-
- E(f) = <e|f> (50)
-
- for unique vector e in F. There exists a unique em such that
-
- f(m) = <em|f> (51) (p.32)
-
- But f(m) = <Km|f>, therefore
-
- em = Km = K(m,m) (52)
-
- also
-
- K(m,m') = <em|em'> (53)
-
- F has a reproducing Kernel if and only if all the evaluation maps are
- bounded.
-
- In applications M has more structure. M can be a Lie group G. M can be a
- "homogeneous space" of Lie group G in which each g in G acts on homogeneous
- space M as an invertible transformation preserving all other structure of M
- like continuity or differentiability. These actions form the Lie group G.
-
- Right action is
-
- m -> mg (54)
-
- Tg:F -> F (55)
-
- (Tgf)(m) = f(mg) (56)
-
- is a "representation" of G on F, since
-
- TgTg' = Tgg' (57)
-
- But f(m) = <Km|f> implies
-
- Kmg = Tg*Km (58)
-
- therefore
-
- K(m'g,mg) = <Km'g|Kmg> = <Km'|TgTg*|Km> (59)
-
- *Therefore, the reproducing kernel K is invariant under the action of G if
- and only if all the operators Tg are unitary, i.e. the representation g->Tg
- is unitary.
-
- Sarfatti problem. In standard quantum mechanics (SQM) what is the group G
- under which K is invariant in order that locally observable probabilities
- (for the complete set of alternatives for part of an entangled nonlocal
- whole) add up to 1? What if G is mutable? Suppose we go to a sub-group H?
- Suppose G is a sub-group of some super-Group? Is this pointing not only to
- something meaningful - but something profound about the relative nature of
- probability in quantum mechanics?
-
- to be continued
-