Alphabeta Math
LemmaStatement: AI-adaptedProof: AI-adaptedPipeline-generatedprecheck pass
How statement and proof provenance work

The first chip identifies the source of the statement or construction; the second identifies the source of its local proof or verification.

  • Literature-sourced: the exact statement appears in a cited source; only wording and notation differ.
  • AI-adapted: a semantically identical restatement of literature-sourced material, modulo indexing, notation, and boundary cases adopted by the library.
  • AI-generated: a genuinely novel statement formulated by AI, with no source for the claim itself.

These labels describe origin, not correctness: citations and verification chips remain separate evidence.

The non-selective Lueders map and the discarded-outcome ensemble

Statement

Assume the Axiom of Countable Choice (The Axiom of Countable Choice (ACω)). Let H be a nonzero complex Hilbert space (Hilbert space), let (Pa)a∈A be a finite or countably infinite family of pairwise orthogonal orthogonal projections on H (The Hilbert orthogonal projection onto a closed subspace) whose strong sum is the identity, ∑a∈APax=x(x∈H), let ρ be a density operator (Density operators and scalar spectral probability clause (a)), and put pa:=tr⁡(ρPa). Define the non-selective Lueders map Φ(ρ):=∑a∈APaρPa. Then:

(a) the series converges in trace norm; Φ(ρ) is a positive trace-class operator with tr⁡Φ(ρ)=1, hence a density operator, and Φ(ρ)=∑aPaΦ(ρ)Pa;

(b) for every b, PbΦ(ρ)=Φ(ρ)Pb=PbΦ(ρ)Pb=PbρPb; in particular Φ(ρ) commutes with every Pb;

(c) discarded-outcome ensemble. Φ(ρ)=∑a:pa>0paρa, where ρa=PaρPa/pa is the selective update of The selective Lueders update by an orthogonal projection; the series converges in trace norm;

(d) if B∈B(H) is bounded and PaB=BPa for every a, then tr⁡(Φ(ρ)B)=tr⁡(ρB);

(e) if ψ=αu+βv with u,v orthonormal, u∈ran⁡P1, v∈ran⁡P2 for distinct labels 1≠2, and ρ=Tψ,ψ, then Φ(ρ)=∣α∣2Tu,u+∣β∣2Tv,v, and tr⁡(Φ(ρ)(Tu,v+Tv,u))=0,tr⁡(ρ (Tu,v+Tv,u))=2Re⁡(αβ‾);

(f) if ρ=Tψ,ψ for a nonzero ψ then Φ(ρ)=∑a:Paψ≠0∥Paψ∥2 Tν(Paψ),ν(Paψ) where ν:H∖{0}→H is the normalization map ν(x):=x/∥x∥, the series converging in trace norm.

No physical interpretation of the map is asserted.

Facts & Assumptions

Given: A nonzero complex Hilbert space H; pairwise orthogonal projections (Pa)a∈A with strong sum I; a density operator ρ; probabilities pa=tr⁡(ρPa); a bounded operator B commuting with every Pa; a vector ψ=αu+βv.

[F1]

Countable Kraus normalization. If (Ma) is a finite or countable family of bounded operators with ∑aMa∗Ma→I strongly, then ∑a∥Max∥2=∥x∥2 for every x; for every positive trace-class ρ the series ∑aMaρMa∗ converges in trace norm to a positive trace-class operator with trace tr⁡ρ; the map extends uniquely to a trace-preserving linear positive map of the trace-class operators, and it maps density operators to density operators (Countable Kraus normalization and trace-norm convergence clauses (a)-(c)).

[F2]

Orthogonal projections and orthogonality. Each Pa satisfies Pa2=Pa=Pa∗ and PaPb=0 for a≠b; ∥Pa∥≤1; the hypothesis of the Statement is that the finite partial sums ∑a≤NPa converge strongly to I (The Hilbert orthogonal projection onto a closed subspace).

[F3]

Zero-probability summands vanish. If pa=tr⁡(ρPa)=0, then PaρPa=0. Indeed, using the expansion [F3'] and trace-norm continuity of left and right multiplication by Pa, PaρPa=∑nwnTPaun,Paun with trace ∑nwn∥Paun∥2=tr⁡(ρPa)=0 (Rank-one operator calculus clauses (b) and (d)); every summand is nonnegative, so each summand vanishes and the series is 0.

[F3'] Expansion of a density operator. ρ=∑nwnTun,un with (un) a finite or countable orthonormal family and wn≥0, ∑nwn=1 (Density operators and scalar spectral probability clause (a)).

[F4]

Rank-one calculus. Tu,vx=⟨x,v⟩u is trace class with tr⁡(Tu,v)=⟨u,v⟩; for bounded B, BTu,v=TBu,v and Tu,vB=Tu,B∗v (Rank-one operator calculus clauses (b)-(d)). For every x,y∈H, Tx,x∗=Tx,x and ⟨Tx,xy,y⟩=∣⟨y,x⟩∣2≥0, so Tx,x is positive by Self-adjoint, positive, unitary and normal operators.

[F5]

Trace-class ideal and cyclicity. For bounded B and trace-class T, BT and TB are trace class with tr⁡(BT)=tr⁡(TB) (Trace class is a two sided Banach operator ideal, Nuclear bounded cyclicity for trace-class operators clause (b)).

[F6]

Selective update. For a single projection P with tr⁡(ρP)>0 the operator ρP=PρP/tr⁡(ρP) is a density operator with PρPP=ρP and QρPQ=0 for every projection Q with QP=0 (The selective Lueders update by an orthogonal projection clauses (a)-(c)).

Proof

1.1givenF1F2algebra

[The map is the projection Kraus map] For each a, Pa∗Pa=Pa2=Pa by [F2], so ∑a≤NPa∗Pa=∑a≤NPa→I strongly: the hypothesis of [F1] holds with Ma=Pa. Therefore Φ(ρ)=∑aPaρPa converges in trace norm, is positive trace class with tr⁡Φ(ρ)=tr⁡ρ=1, and is a density operator. This proves the first part of (a).

2.1step 1.1F1F2F5algebra

[The compressed identity] Fix b. Left and right multiplication by the bounded operator Pb are continuous in trace norm, since ∥PbT∥1≤∥Pb∥ ∥T∥1 and ∥TPb∥1≤∥Pb∥ ∥T∥1 for trace-class T (Trace class is a two sided Banach operator ideal); applying this to the trace-norm convergent series of step 1.1 gives PbΦ(ρ)=∑aPbPaρPa and Φ(ρ)Pb=∑aPaρPaPb. Since PbPa=0 for a≠b and Pb2=Pb [F2], all summands vanish except a=b, and that summand is PbρPb; hence PbΦ(ρ)=PbρPb=Φ(ρ)Pb and therefore PbΦ(ρ)Pb=PbρPb. Summing the identity PbΦ(ρ)Pb=PbρPb over b and using step 1.1 once more gives Φ(ρ)=∑bPbΦ(ρ)Pb. This proves (a) and (b).

2.2step 1.1F3F6algebra

[The discarded-outcome mixture] For a label a with pa>0, [F6] gives PaρPa=paρa with ρa a density operator, and [F3] gives PaρPa=0 for a label with pa=0. Hence the summands with pa=0 vanish and, reindexing the trace-norm convergent series of step 1.1 over the subfamily {a:pa>0}, Φ(ρ)=∑a:pa>0paρa. This is (c).

2.3step 1.1F1F5algebra

[Invariance under compatible bounded quantities] Let B be bounded with PaB=BPa for every a. Then PaBρPa=BPaρPa and, using the trace-norm continuity of left multiplication by B together with step 1.1, BΦ(ρ)=∑aPa(Bρ)Pa. The right-hand side is the Kraus map of [F1] evaluated at the trace-class operator Bρ, so tr⁡(BΦ(ρ))=tr⁡Φ(Bρ)=tr⁡(Bρ) by trace preservation; since tr⁡(Φ(ρ)B)=tr⁡(BΦ(ρ)) [F5], this gives tr⁡(Φ(ρ)B)=tr⁡(ρB). This is (d).

3.1step 2.2F2F4algebra

[Coherence destruction] Let ψ=αu+βv with u,v orthonormal, u∈ran⁡P1 and v∈ran⁡P2, 1≠2, and ρ=Tψ,ψ. For a∉{1,2} one has Pau=PaP1u=0 and Pav=PaP2v=0 by [F2], so Paψ=0; and P1ψ=αu, P2ψ=βv. Hence the series has only two nonzero summands, and [F4] gives P1ρP1=TP1ψ,P1ψ=∣α∣2Tu,u and P2ρP2=∣β∣2Tv,v, so Φ(ρ)=∣α∣2Tu,u+∣β∣2Tv,v. For the pairing with Tu,v: tr⁡(Tu,uTu,v)=⟨Tu,uu,v⟩=⟨u,v⟩=0 and tr⁡(Tv,vTu,v)=⟨Tv,vu,v⟩=⟨u,v⟩⟨v,v⟩=0, and conjugating gives the same vanishing for Tv,u; hence tr⁡(Φ(ρ)(Tu,v+Tv,u))=0. For the coherent state, tr⁡(ρTu,v)=⟨Tψ,ψu,v⟩=⟨u,ψ⟩⟨ψ,v⟩=α‾β and tr⁡(ρTv,u)=⟨u,ψ⟩⟨ψ,v⟩‾=αβ‾, so the cross term is 2Re⁡(αβ‾). This is (e).

4.1step 1.1F4algebra∎

[The pure-state formula] Let ρ=Tψ,ψ with ψ≠0. By step 1.1 with ρ=Tψ,ψ and the rank-one rules [F4], PaρPa=TPaψ,Paψ, and TPaψ,Paψ=∥Paψ∥2Tν(Paψ),ν(Paψ) when Paψ≠0, while both sides vanish when Paψ=0; the trace-norm convergence of the resulting series is that of step 1.1. This is (f).

Depends on

Dependency tree · 2 levels

Sources