Simultaneous Classical Identification

• Simultaneous classical identification is a paradigm where a single measurement tests if a candidate message was sent, merging classical hypothesis testing with quantum measurement constraints.
• The approach employs a unified decoder where all binary hypothesis tests share one observation, ensuring that the identification capacity does not exceed the classical transmission capacity.
• Applications in CCQ channels and qubit depolarizing channels demonstrate that even with doubly exponential codebook growth, the operational rate remains equal to the classical capacity.

Searching arXiv for recent and foundational papers on simultaneous classical identification. Simultaneous classical identification is a variant of message identification in which a receiver does not decode the transmitted message in the Shannon sense, but instead tests whether a specific candidate identity was sent. In the simultaneous form, all such tests must be implemented through a single measurement or decoding structure rather than a separate test for each candidate. This restriction is operationally central in quantum settings, where distinct measurements cannot generally be applied to the same received state without disturbance. For classical–quantum multiple-access channels, the simultaneous identification capacity region coincides with the classical message-transmission capacity region (Boche et al., 2018). For the qubit depolarizing channel under complete product measurements, the simultaneous classical identification capacity likewise equals the classical capacity, with exact single-letter value $1-h(p/2)$ (Ye et al., 31 Mar 2026). Related work also connects identification to feedback-assisted sensing over discrete memoryless channels, where doubly exponential codebook growth remains compatible with a distortion constraint (Labidi et al., 2023).

1. Concept and formal objective

Message identification, as formalized by Rudolf Ahlswede and Gunter Dueck, replaces the decoding question “Which message was sent?” with a family of binary hypothesis tests of the form “Was message $m$ sent?” In this paradigm, randomized encoding permits code sizes that grow doubly exponentially with blocklength, so the operational rate is measured through $(1/k)\log\log M$ rather than $(1/k)\log M$ (Boche et al., 2018). The same scaling appears in discrete memoryless channels with feedback-assisted sensing, where an achievable identification rate-distortion pair $(R,D)$ supports $N=2^{2^{nR}}$ identities (Labidi et al., 2023).

The qualifier “simultaneous” refers to the requirement that all identification decisions be derivable from a single observation interface. In the classical–quantum setting, this is expressed through one POVM together with message-dependent index sets; the identifiers $I_{mn}$ are obtained by coarse-graining the outcomes of that single POVM (Boche et al., 2018). In the qubit depolarizing setting with complete product measurements, the single measurement is a refining POVM $$\{E_{x^n}=|\Psi_{x^n}\rangle\langle\Psi_{x^n}|\}_{x^n\in\{0,1\}^n}$$, and each decoder is a sum over a decision region $I_i\subset\{0,1\}^n$ (Ye et al., 31 Mar 2026).

A useful operational interpretation is that simultaneous identification imposes compatibility among all yes/no tests. In quantum channels this compatibility is not merely aesthetic; it is required to respect measurement disturbance (Boche et al., 2018). In classical models, the same formalism can be realized through a single decoder output and a family of decision regions (Labidi et al., 2023).

2. Code structure and error criteria

For a discrete, memoryless classical–classical–quantum multiple-access channel

$W:X\times Y\to\mathcal S(H),$

with two independent classical senders and one quantum receiver, an identification code of blocklength $m$0 consists of randomized encoders $m$1, $m$2, and identifiers $m$3 with $m$4 on $m$5 (Boche et al., 2018). The two standard error criteria are:

• Type I, or missed identification:

$m$6

• Type II, or false acceptance:

$m$7

The rates $m$8 are achievable if for all $m$9 and all large $(1/k)\log\log M$0, there exists an ID-code satisfying

$(1/k)\log\log M$1

(Boche et al., 2018). Because $(1/k)\log\log M$2 and $(1/k)\log\log M$3 scale linearly in $(1/k)\log\log M$4, the message-set sizes are approximately

$(1/k)\log\log M$5

(Boche et al., 2018).

In the qubit depolarizing case, a length-$(1/k)\log\log M$6 simultaneous classical-ID code with complete product measurement consists of encodings $(1/k)\log\log M$7, a single refining POVM in a product basis, and decision regions $(1/k)\log\log M$8 such that

$(1/k)\log\log M$9

(Ye et al., 31 Mar 2026). The corresponding capacity is defined through $(1/k)\log M$0 (Ye et al., 31 Mar 2026).

3. Simultaneity as a measurement constraint

In the multiple-access CCQ model, a code is simultaneous if there exists a single POVM

$(1/k)\log M$1

and index sets $(1/k)\log M$2, $(1/k)\log M$3 such that

$(1/k)\log M$4

(Boche et al., 2018). The receiver performs one measurement and then tests whether the outcome satisfies $(1/k)\log M$5 and $(1/k)\log M$6. This is the precise formalization of simultaneous identification in that setting.

The necessity of simultaneity is especially transparent in quantum models. The received state cannot, in general, support an independently chosen measurement for each message pair, because measurement disturbs the state (Boche et al., 2018). The restriction therefore encodes a physical consistency requirement rather than merely a decoder simplification.

A related but distinct notion appears in work on multipartite classical states and non-disruptive local state identification (NDLID). There, a state is fully classical when each party can perform a projective measurement to identify a locally held state without disturbing the global state, simultaneously for every party (Chen et al., 2010). This is not the same problem as channel identification, but it exhibits a structurally similar concern: information extraction is constrained by disturbance. The comparison suggests that “simultaneous” identification often marks the boundary between formally available tests and physically co-implementable tests.

4. Capacity theorems and single-letter characterizations

The central theorem for the CCQ multiple-access channel states that the simultaneous ID-capacity region coincides with the classical message-transmission capacity region: $(1/k)\log M$7 (Boche et al., 2018). The transmission region is

$(1/k)\log M$8

where

$(1/k)\log M$9

(Boche et al., 2018). The entropic quantities are given by

$(R,D)$0

and

$(R,D)$1

with $(R,D)$2 the von Neumann entropy (Boche et al., 2018).

For the qubit depolarizing channel under complete product measurements, the corresponding exact formula is

$(R,D)$3

(Ye et al., 31 Mar 2026). The lower bound is obtained from the general inequality

$(R,D)$4

and King’s formula for the classical capacity of the depolarizing channel gives $(R,D)$5 (Ye et al., 31 Mar 2026). The converse matches this bound under the stated measurement restriction, yielding equality.

These results collectively support a general pattern: once simultaneity is imposed, identification does not enlarge the asymptotic rate region beyond the message-transmission region, even though the codebook cardinality is doubly exponential (Boche et al., 2018, Ye et al., 31 Mar 2026). A plausible implication is that the combinatorial abundance of identification messages is compensated by the stronger compatibility constraints on the decoder.

5. Proof strategies

The achievability proof for CCQ multiple-access channels proceeds through what the paper terms the “transformator lemma.” One starts from a classical transmission code of blocklength $(R,D)$6 with rates in $(R,D)$7 and average error tending to zero, appends a short maximal-error code of length $(R,D)$8 with exponentially smaller rates, and constructs identification distributions on total length $(R,D)$9 by concatenating transmission codewords and randomness codewords (Boche et al., 2018). The identifiers are defined as direct sums $N=2^{2^{nR}}$0 according to code indices and random map families, and Chernoff–Hoeffding bounds control the random choice of mappings so that both $N=2^{2^{nR}}$1 and $N=2^{2^{nR}}$2 tend to zero (Boche et al., 2018).

The converse for CCQ channels begins with a simultaneous ID-code having doubly exponential $N=2^{2^{nR}}$3 and small $N=2^{2^{nR}}$4. The argument refines the POVM to bound the trace distance between output distributions of distinct codewords, then invokes a classical-channel resolvability argument, specifically Steinberg’s lemma, to extract input distributions achieving large mutual information $N=2^{2^{nR}}$5 and $N=2^{2^{nR}}$6. These rates must lie in the region defining $N=2^{2^{nR}}$7, which excludes identification rates outside the classical transmission region (Boche et al., 2018).

For the qubit depolarizing channel, the converse relies on two observations. First, a complete product measurement in a product basis followed by ID-decoding is equivalent to measuring in that basis and then checking whether the observed string $N=2^{2^{nR}}$8 lies in a decision region $N=2^{2^{nR}}$9. Second, in that basis the action of $I_{mn}$0 reduces to an $I_{mn}$1-fold binary symmetric channel with crossover $I_{mn}$2 on the diagonal of the input: $I_{mn}$3 (Ye et al., 31 Mar 2026). A classical soft-covering argument then shows that if the number of types exceeds the covering threshold governed by $I_{mn}$4, two outputs must become too close in total variation, contradicting the separation required for identification (Ye et al., 31 Mar 2026). This yields the bound

$I_{mn}$5

whenever $I_{mn}$6 (Ye et al., 31 Mar 2026).

The feedback-assisted sensing model uses a different achievability architecture. The encoder repeatedly sends a fixed symbol $I_{mn}$7, and noiseless feedback gives both encoder and receiver access to the same random output sequence $I_{mn}$8. Typical-set arguments provide a high-probability set $I_{mn}$9 of size approximately $$\{E_{x^n}=|\Psi_{x^n}\rangle\langle\Psi_{x^n}|\}_{x^n\in\{0,1\}^n}$$0 on which the induced distribution is nearly uniform; random coloring functions $$\{E_{x^n}=|\Psi_{x^n}\rangle\langle\Psi_{x^n}|\}_{x^n\in\{0,1\}^n}$$1 and a short deterministic transmission code of length $$\{E_{x^n}=|\Psi_{x^n}\rangle\langle\Psi_{x^n}|\}_{x^n\in\{0,1\}^n}$$2 then realize identification (Labidi et al., 2023). This construction shows how identification can be embedded into a broader joint task while preserving doubly exponential scaling.

6. Relations to classical identification, sensing, and state structure

In the purely classical two-sender MAC setting, Ahlswede–Dueck showed that the simultaneous identification capacity region equals the message-transmission region under a suitably defined simultaneous identification code, and the CCQ result extends that equality to the classical–quantum setting (Boche et al., 2018). The extension is significant because the quantum receiver introduces a nontrivial measurement compatibility constraint, yet the ultimate single-letter region remains unchanged.

The joint identification-and-sensing model generalizes the classical identification paradigm to a state-dependent DMC with i.i.d. states and strictly causal feedback (Labidi et al., 2023). There, the sender transmits an ID message while simultaneously estimating the state sequence under a blockwise average distortion constraint

$$\{E_{x^n}=|\Psi_{x^n}\rangle\langle\Psi_{x^n}|\}_{x^n\in\{0,1\}^n}$$3

(Labidi et al., 2023). The paper proves the lower bound

$$\{E_{x^n}=|\Psi_{x^n}\rangle\langle\Psi_{x^n}|\}_{x^n\in\{0,1\}^n}$$4

where $$\{E_{x^n}=|\Psi_{x^n}\rangle\langle\Psi_{x^n}|\}_{x^n\in\{0,1\}^n}$$5 is the per-symbol minimum-distortion function and $$\{E_{x^n}=|\Psi_{x^n}\rangle\langle\Psi_{x^n}|\}_{x^n\in\{0,1\}^n}$$6 (Labidi et al., 2023). This suggests that simultaneous classical identification is compatible with auxiliary sensing functionality when the channel can also generate common randomness through feedback.

The work on generalized-classical and classical multipartite states provides an operationally adjacent notion of simultaneous identifiability at the state level rather than the channel level. Fully classical states are those for which each party can identify its local component by projective measurement without disturbing the global state; fully generalized-classical states admit the analogous task with general POVMs (Chen et al., 2010). A plausible implication is that simultaneous classical identification in channels and NDLID in states share a common structural theme: both isolate those settings in which many logically distinct binary tests can be consolidated into a single physically admissible observation procedure.

7. Limits, misconceptions, and scope

A common misconception is that doubly exponential codebook growth should imply an identification rate strictly larger than the underlying transmission capacity. The cited results do not support that conclusion under simultaneous decoding constraints. In the CCQ multiple-access setting, the simultaneous identification capacity region is exactly $$\{E_{x^n}=|\Psi_{x^n}\rangle\langle\Psi_{x^n}|\}_{x^n\in\{0,1\}^n}$$7, not a larger region (Boche et al., 2018). For the qubit depolarizing channel with complete product measurements, the simultaneous classical identification capacity is exactly $$\{E_{x^n}=|\Psi_{x^n}\rangle\langle\Psi_{x^n}|\}_{x^n\in\{0,1\}^n}$$8, again equal to the classical capacity (Ye et al., 31 Mar 2026).

Another possible misunderstanding is that quantum channels necessarily provide a quantum-specific boost in simultaneous identification. The CCQ multiple-access result explicitly states that no such boost occurs once one enforces the simultaneous measurement constraint (Boche et al., 2018). Likewise, for the depolarizing channel under complete product measurements, the converse bound vanishes as $$\{E_{x^n}=|\Psi_{x^n}\rangle\langle\Psi_{x^n}|\}_{x^n\in\{0,1\}^n}$$9, so the capacity correctly tends to zero in the completely noisy limit: $I_i\subset\{0,1\}^n$0 (Ye et al., 31 Mar 2026). This corrects the behavior of previously known converse bounds that remained strictly positive even for a completely noisy channel (Ye et al., 31 Mar 2026).

At the same time, the exact equality results are model-dependent. The depolarizing-channel theorem is stated under the constraint of complete product measurements (Ye et al., 31 Mar 2026), and the sensing result currently provides a lower bound rather than a complete capacity-distortion characterization in the form summarized here (Labidi et al., 2023). Accordingly, the strongest established conclusions concern simultaneous identification under explicitly specified decoding or measurement architectures.

Taken together, these works place simultaneous classical identification at a junction of identification theory, multiuser information theory, quantum measurement theory, and feedback-assisted communication. The defining phenomenon is the coexistence of doubly exponential message families with single-measurement or single-decoder realizability; the dominant conclusion, across the best-characterized models, is that this dramatic combinatorial enlargement does not alter the governing single-letter information bounds (Boche et al., 2018, Ye et al., 31 Mar 2026).