{"id":"404f375e-50c1-4c58-9b09-771acbc953d6","arxiv_id":"2504.15025","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"Quantum pseudoresources, including pseudoentanglement, imply EPFI pairs, which yield quantum commitments, oblivious transfer, and secure multiparty computation.","lead":"The paper shows that quantum states that look resource-rich but cannot be efficiently distinguished from resource-poor ones can be turned into cryptographic building blocks such as commitments. This suggests that pseudoentanglement and other quantum pseudoresources are enough for secure computation, not just curiosities.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Worst-case resource gap may not hold for known pseudoentanglement constructions, so the central implication may not cover the standard notion.","rationale":"The paper's formal chain—pseudoresource implies EPFI, EPFI implies commitments, commitments imply EFI and thus OT/MPC—is internally coherent, and the resource-gap-to-trace-distance inversion using Winter's inequality is sound once the division by kappa and the choice of constants are read correctly. The reader's verdict of CONDITIONAL is appropriate. I considered two other defects: Theorem 2.13's exponential bound on amplified trace distance is stated with exponent n*Delta/2, which is false for small Delta (two biased coins with Delta=0.01 need roughly n*Delta^2 copies, not n*Delta); this makes the proof of Theorem 3.3 technically invalid as written, though a correct Chernoff bound with exponent n*Delta^2/2 restores the argument. I also noted the abstract's 'equivalent' claim is only proven in the direction EPFI implies commitments, though EFI implies EPFI is trivial, so the equivalence is recoverable. Neither of these threatens the central construction as much as the worst-case gap issue, because they are readily fixable. The worst-case issue determines the theorem's scope: if known pseudoentanglement constructions only provide a gap with overwhelming probability, then the paper's theorems do not apply to them, and the paper's own claims that they do are unsupported. The fix is a missing averaging or key-restriction argument, not a change to the main construction, so the verdict should remain CONDITIONAL rather than REJECT.","tokens_in":18836,"tokens_out":18978,"duration_ms":189749,"concrete_test":"Examine the construction in [ABF+23] (and, if needed, [ABV23] and [GE24]) and determine whether the high-entanglement family is defined so that every index/key in the family yields a state with entanglement at least the claimed threshold. Concretely: list the random choices used to sample a high-entanglement state; check whether any choice with non-negligible probability produces a state below the threshold. If such choices exist, test whether restricting the key space to 'good' random choices preserves efficient sampling and computational indistinguishability (the failure probability is negligible, so rejection sampling should work in expectation, but the paper must supply this argument). If no such restriction is possible, Theorem 5.2 does not apply to that construction and the claim that pseudoentanglement implies EPFI pairs is limited to the paper's strengthened definition.","verdict_should_be":"UNCHANGED","load_bearing_attack":"Definitions 4.2, 5.1, and 5.4 require the resource/entanglement gap to hold for every pair of keys (k,k'). The EPFI definition (3.2) and the commitment binding condition (Eq. 5) likewise quantify over all pairs. The paper asserts in Section 5.1 that 'every known construction of pseudoentanglement from pure state ensembles' satisfies the required gap, and in Section 5.2 that any gap in the computational measures of [ABV23]/[GE24] transfers to the regularized relative entropy of entanglement. However, pseudoentanglement constructions such as [ABF+23] are typically distributional: the high-entanglement ensemble is sampled from a distribution (e.g., Haar-random states) and the stated entanglement lower bound holds only with overwhelming probability over the sampler's randomness, not for every key in the index set. If exceptional keys exist with negligible probability, the worst-case gap fails, and with it the pairwise trace-distance lower bound needed for EPFI and for honest statistical binding, because the binding proof must hold for all k,k'. The paper provides no averaging or key-restriction argument to convert a high-probability gap into the all-pairs gap its theorems require. Hence the headline result 'pseudoresources imply cryptography' is rigorously established only for a new, stronger all-pairs definition, and its applicability to previously studied pseudoentanglement is unsubstantiated.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper introduces EPFI pairs, a family-indexed generalization of EFI pairs in which every element of one family is statistically far from every element of the other, while the two families are computationally indistinguishable. It proves (Theorem 3.3) that EPFI pairs imply statistically binding, computationally hiding canonical quantum commitments, thereby inheriting known consequences such as oblivious transfer and multiparty computation. The main technical claim (Theorem 4.3) is that any 'η-gap pseudoresource'—two efficiently generated, computationally indistinguishable families of states with a large gap in relative entropy of resource—yields EPFI pairs, provided the resource gap is at least 2+1/poly(n) and the maximum resource variation κ is polylog(d). The paper then specializes to entanglement, giving results for pure-state pseudoentanglement (Theorem 5.2) and mixed-state pseudoentanglement (Corollary 5.5), and proposes a new functionality called computationally locked entanglement. The paper claims that every known pure-state pseudoentanglement construction satisfies the required gap.","tokens_in":18958,"tokens_out":15606,"duration_ms":133077,"significance":"If the results hold as stated, the paper would establish a general bridge from quantum resource theories (entanglement, magic, coherence) to quantum cryptography, complementing known minimal-assumption results for EFI pairs. The EPFI-to-commitment reduction is a clean and useful generalization of the EFI construction, and the paper is careful to spell out the commitment scheme and its security proof. The asymptotic-continuity-based approach to turning resource gaps into trace-distance gaps is elegant. However, the current formulation relies on a worst-case (all-pairs) gap condition that is not obviously satisfied by existing pseudoentanglement constructions, and the proof of the central pseudoresource theorem contains an incorrect inversion of Winter's inequality. These issues currently limit the scope of the claimed implication.","major_comments":[{"comment":"The proof step 'it follows from Equation (3) that, for every k,k′∈{0,1}^κ, Δ(ψ_k,φ_k′) ≥ |R_rel(ψ_k)-R_rel(φ_k′)| - 2/κ' is not a consequence of Lemma 2.16. The lemma gives, for any ρ,σ with Δ(ρ,σ) ≤ ε, the upper bound |R_rel(ρ)-R_rel(σ)| ≤ εκ + (1+ε)h(ε/(1+ε)). Inverting this requires choosing ε such that εκ + (1+ε)h(ε/(1+ε)) < η, and the resulting lower bound on Δ does not take the stated form. As written, the right-hand side can exceed 1 (e.g., for η=2 and κ=10 it is 1.8), so the inequality cannot hold for a trace distance. This invalidates the proof of the central theorem; the authors must provide a correct inversion and verify that the conditions η ≥ 2+1/poly(n) and κ = polylog(d) indeed imply Δ ≥ Ω(1/poly).","section":"Section 4.2, Theorem 4.3"},{"comment":"The 'for all k,k′' entanglement gap in Definition 5.1 is a worst-case condition that standard pseudoentanglement constructions do not satisfy. In the ABF+23-style construction, the high-entanglement ensemble is a distribution over (near-)Haar-random states, whose entanglement is at least n/2 - O(1) only with probability 1 - negl; the key set may include exceptional states with low entanglement. Since the EPFI farness condition (Definition 3.2) and the commitment binding condition (Eq. (5)) are also quantified over all keys, the theorem does not apply to these average-case constructions. The paper's assertion in Section 5.1 that 'every known construction of pseudoentanglement from pure state ensembles ... exhibits an entanglement entropy gap of at least 1/2e + 1/poly(n)' is therefore unsubstantiated unless the authors provide an argument that restricts the key space to good keys (or otherwise converts the high-probability gap into an all-pairs gap) while preserving computational indistinguishability. The same concern applies to the mixed-state pseudoentanglement definitions in Definitions 5.4 and Corollary 5.5.","section":"Section 5.1, Definition 5.1 and Theorem 5.2"}],"minor_comments":[{"comment":"The informal description says the resource gap holds for 'states sampled from each family', but Definition 4.2 requires the gap for all key pairs; this discrepancy should be resolved because it affects the scope of the results.","section":"Introduction, Informal Definition 1 vs Definition 4.2"},{"comment":"In Definition 2.9, the condition '∀k∈N+' should be '∀λ∈N+'; the same typo appears in Definition 2.10.","section":"Section 2.3, Definitions 2.9 and 2.10"},{"comment":"The constant 'c' in the Fannes bound is actually c(Δ); please write it as c(Δ) and use the bound c(Δ) ≤ 1/(2e) explicitly to derive the stated lower bound.","section":"Section 5.1, proof of Theorem 5.2"},{"comment":"The statement 'ˆE^ε_D < ˆE^ε_C' should be '≤', and the proof would benefit from a more explicit description of how the distillation algorithm yields a distinguisher between the two families.","section":"Section 5.3, Lemma 5.7"},{"comment":"There are several typographical errors, e.g., 'statistical biding' in Section 2.6, 'ω(log(n)))' in Section 5.1, and 'ca be extended' in Section 3; a careful proofreading would improve readability.","section":"Throughout"}],"recommendation":"major_revision","confidential_remarks":"The paper's core idea is appealing and the EPFI-to-commitment reduction is solid, but the two issues above (the incorrect inversion in Theorem 4.3 and the unproven coverage of average-case pseudoentanglement constructions) are load-bearing. Both seem correctable: the first is likely a typographical error in the inversion, and the second can be addressed by an explicit key-restriction argument or by revising the scope claims. I therefore support major revision rather than rejection. No concerns about citation practices; the only self-citation (GLSV21) is to a published CRYPTO paper used for a standard implication."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The paper introduces EPFI pairs and shows that a pseudoresource gap in relative entropy implies them, and that EPFI pairs imply canonical quantum commitments. The reductions are clean and the EPFI notion is a natural generalization of EFI. The main caveat is that the gap is required for every pair of keys, and the paper asserts without proof that known pseudoentanglement constructions satisfy that worst-case condition. If that assertion is not true, the headline claim only covers a new, stronger definition.\n\nWhat is genuinely new: the EPFI primitive; the template that converts an asymptotic-continuity bound into a trace-distance lower bound; the clean EPFI-to-commitment proof via Uhlmann's theorem; and the 'computationally locked entanglement' proposal. The paper is careful about the resource-theoretic setup, and the reductions are honest rather than fitted to the conclusion.\n\nSoft spots. First, the inversion of Winter's inequality is typeset incorrectly in Theorem 4.3; the displayed inequality Delta >= |gap| - 2/kappa would force trace distance above 1. The intended bound divides by kappa and follows by contrapositive. The same problem appears in Theorem 5.2 with Fannes' inequality. These are fixable but must be corrected. Second, the worst-case gap. Definitions 4.2, 5.1, and 5.4 quantify over all key pairs, and the EPFI farness and commitment binding require all-pairs trace distance. Many pseudoentangled constructions in the literature guarantee the gap with overwhelming probability over the key, not for every key. The paper's assertion in Section 5.1 that 'every known construction' satisfies the worst-case gap is unsubstantiated. If it fails, the main theorem does not apply to those definitions as stated. Third, the abstract says EPFI pairs are equivalent to commitments; this is true via the chain EPFI -> commitments -> EFI -> EPFI, but the body only explicitly proves the first arrow, and the EFI-to-EPFI step is just the observation that EFI is the one-key case. Worth spelling out.\n\nThis is not a takedown. The core logic under the paper's own definitions is sound, and the soft spots are typos plus a missing bridge to prior work. It is a solid conceptual contribution for quantum cryptography and resource theories, and it deserves serious peer review. The authors should fix the inversions and either prove that known constructions satisfy the worst-case gap or weaken the theorem to a high-probability gap with a proper averaging argument.","headline":"Clean reduction from an all-keys pseudoresource gap to commitments via a new EPFI primitive, but the claimed coverage of known pseudoentanglement definitions is not substantiated.","tokens_in":19607,"tokens_out":8707,"would_cite":true,"duration_ms":81607,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":["03.67.Dd","03.67.Mn"],"model":"deepseek-v4-flash","headline":"Quantum pseudoresources—state families that hide a resource gap from efficient observers—are enough to build commitments, oblivious transfer, and secure multiparty computation.","keywords":["quantum pseudoresources","EPFI pairs","quantum commitments","pseudoentanglement","EFI pairs","relative entropy of resource","computational indistinguishability","oblivious transfer"],"falsifier":"Take any candidate $\\eta$-gap pseudoresource satisfying the hypotheses of the main theorem and compute $\\min_{k,k'} \\Delta(\\psi_k, \\phi_{k'})$ for the constructed EPFI pair. The theorem predicts this minimum is $\\Omega(1/\\mathrm{poly}(n))$; a single key pair with negligible trace distance, or a computationally indistinguishable family with only average-case resource gap containing an exceptional pair with near-equal resource, would falsify the conclusion.","tokens_in":18504,"feed_emoji":"🔐","tokens_out":13442,"duration_ms":112800,"temperature":0.7,"pith_summary":"The paper aims to show that a single general phenomenon in quantum resource theories—two efficiently generable state families that are computationally indistinguishable yet have a large gap in some resource, from entanglement to magic to coherence—is already enough to build quantum cryptography. It introduces EPFI pairs, keyed families of states that are pairwise far in trace distance yet indistinguishable, and proves these yield statistically binding, computationally hiding canonical quantum commitments. The paper then shows that an $\\eta$-gap pseudoresource, measured by relative entropy of resource with $\\eta \\geq 2 + 1/\\mathrm{poly}(n)$, produces EPFI pairs through a continuity inequality; pure and mixed pseudoentanglement give the same outcome with slightly different thresholds. If this is right, quantum resources, not just one-way functions or pseudorandomness, form a general cryptographic engine.","feed_headline":"Hidden quantum resource gaps can build cryptography","feed_subtitle":"Two indistinguishable state families with a large resource gap yield commitments and secure computation.","key_machinery":"The central object is the relative entropy of resource, $R_{\\mathrm{rel}}(\\rho) = \\min_{\\sigma \\in F} D(\\rho\\|\\sigma)$, measuring the state's distance from the free states of a resource theory. The argument is carried by a continuity bound for this quantity: when two states are $\\epsilon$-close in trace distance, their resource values differ by at most $\\epsilon \\kappa + (1+\\epsilon) h(\\epsilon/(1+\\epsilon))$, where $\\kappa$ is the maximum variation of $R_{\\mathrm{rel}}$ over the state space and $h$ is the binary entropy. Arranging the resource gap $\\eta \\geq 2 + 1/\\mathrm{poly}(n)$ with $\\kappa = \\mathrm{polylog}(d)$ forces every cross-pair of pseudoresource states to be $\\Omega(1/\\mathrm{poly}(n))$ far in trace distance. The EPFI-to-commitment step is carried by amplification of many copies plus the purification-fidelity relation: pairwise trace-distance farness becomes a vanishing fidelity gap, giving honest statistical binding, while computational indistinguishability directly gives hiding.","core_discovery":"On its own terms, the central claim is that pseudoresources are cryptographically sufficient: for $\\eta \\geq 2 + 1/\\mathrm{poly}(n)$, an $\\eta$-gapped pseudoresource (with the total variation $\\kappa$ of the relative entropy of resource at most polylogarithmic in the dimension) yields EPFI pairs, and EPFI pairs yield canonical quantum commitments, hence oblivious transfer, secure multiparty computation, and related primitives. For pure-state pseudoentanglement measured by entanglement entropy, the threshold is $\\eta \\geq 1/(2e) + 1/\\mathrm{poly}(n)$, via reduced density matrices and an entropy-continuity inequality; for mixed states measured by regularized relative entropy of entanglement, it is $\\eta \\geq 2 + 1/\\mathrm{poly}(n)$, via the corresponding continuity bound. The same machinery yields a new proposed functionality, computationally locked entanglement, in which high entanglement is hidden from keyless observers but distillable with the key.","pith_inferences":["Not shown in the paper, but a natural next test: can the worst-case per-key gap be relaxed to an average-case gap? The proofs quantify over all key pairs, so a positive result would substantially widen the class of pseudoresources that yield cryptography.","The paper leaves the threshold $\\eta \\geq 2 + 1/\\mathrm{poly}(n)$ tied to the continuity inequality used; a tighter inequality would lower the gap and improve the commitment parameters. That quantitative relationship is an inference from the proof structure, not an additional theorem in the paper.","If the reverse direction holds—EPFI pairs or commitments themselves yield some pseudoresource—then pseudoresources would sit exactly at the minimal-assumption level of quantum cryptography; the paper proves only the forward direction.","Computationally locked entanglement relaxes the requirement that both families be efficiently generated, so it may be constructible from weaker assumptions than full pseudoentanglement; whether it is would clarify which resource-gap notions are essential."],"forward_implications":["Any resource theory whose free states are convex, closed, and contain a full-rank state inherits the result: an $\\eta$-gapped pseudoresource with $\\eta \\geq 2 + 1/\\mathrm{poly}(n)$ and $\\kappa = \\mathrm{polylog}(d)$ yields EPFI pairs and therefore quantum commitments.","All known pure-state pseudoentanglement constructions with an entanglement-entropy gap of at least $1/(2e) + 1/\\mathrm{poly}(n)$ become usable for cryptography, because their reduced density matrices form EPFI pairs.","Mixed-state pseudoentanglement defined through the regularized relative entropy of entanglement, with gap at least $2 + 1/\\mathrm{poly}(n)$, yields EPFI pairs even when the gap is invisible to the computational entanglement measures of earlier definitions.","From EPFI pairs, the paper obtains statistically binding and computationally hiding canonical quantum commitments; by standard reductions these give oblivious transfer, secure multiparty computation, and further quantum cryptographic protocols.","Computationally locked entanglement, if it can be instantiated, supplies a key-controlled entanglement functionality suitable for authenticated teleportation and certified routing in quantum networks, usable polynomially many times."],"supporting_citations":[{"why":"Introduces EFI pairs and proves their equivalence to canonical quantum commitments, the target structure that EPFI pairs are designed to extend.","marker":"[BCQ23]"},{"why":"Defines canonical quantum commitments and shows honest statistical binding suffices, which the EPFI-based commitment uses.","marker":"[Yan22]"},{"why":"Provides the continuity bound for relative entropy that converts the resource gap into a trace-distance lower bound in the pseudoresource theorem.","marker":"[Win16]"},{"why":"Provides the entropy-continuity inequality used to turn an entanglement-entropy gap between pure pseudoentangled states into statistical distance of the reduced EPFI states.","marker":"[Fan73]"},{"why":"Provides the purification-fidelity relation that converts pairwise trace-distance farness into the fidelity bound needed for honest statistical binding.","marker":"[Uhl76]"},{"why":"Introduces the pure-state pseudoentanglement framework whose constructions the paper re-uses to produce EPFI pairs.","marker":"[ABF+23]"},{"why":"Supplies the computational entanglement measures and pseudoentanglement definitions that the mixed-state part of the paper builds on.","marker":"[ABV23]"},{"why":"Established mixed pseudoentanglement as a minimal assumption for EFI pairs, the necessity counterpart to the sufficiency shown here.","marker":"[GE24]"}],"fun_headline_variants":["Pseudoresources unlock quantum cryptography","Resource gaps forge quantum cryptographic keys","Quantum state gaps yield cryptographic power","From hidden gaps to quantum commitments"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"All constructions assume the resource gap holds for every pair of keys—a worst-case gap—and that the resource measure's total variation $\\kappa$ is polylogarithmic in dimension; if either fails, the pairwise trace-distance farness that binding requires is not guaranteed.","fun_headline_variants_meta":{"raw":{"variants":["Pseudoresources unlock quantum cryptography","Resource gaps forge quantum cryptographic keys","Quantum state gaps yield cryptographic power","From hidden gaps to quantum commitments"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000184,"raw_usage":{"total_tokens":1320,"prompt_tokens":946,"completion_tokens":374,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":562,"completion_tokens_details":{"reasoning_tokens":327}},"tokens_in":562,"tokens_out":374,"duration_ms":3541,"temperature":1.0,"reasoning_tokens":327,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-16T11:37:05.422668+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Take any candidate $\\eta$-gap pseudoresource satisfying the hypotheses of the main theorem and compute $\\min_{k,k'} \\Delta(\\psi_k, \\phi_{k'})$ for the constructed EPFI pair. The theorem predicts this minimum is $\\Omega(1/\\mathrm{poly}(n))$; a single key pair with negligible trace distance, or a computationally indistinguishable family with only average-case resource gap containing an exceptional pair with near-equal resource, would falsify the conclusion.","supporting_citations":[],"review_version":1}