{"id":"5c6b2296-3f51-426b-8357-24bcd252d56d","arxiv_id":"2607.19870","paper_version":1,"verdict":"REJECT","confidence":"MODERATE","novelty_score":4.0,"correctness_risk":"high","formal_verification":"none","parameter_count":2,"one_line_summary":"A two-particle walk with motion-only interactions is claimed to require a boundary-memory ancilla for unitarity and to show Page-curve-like entropy, but Eq. 5 is demonstrably non-unitary.","lead":"This paper builds a two-particle quantum walk where attraction and repulsion are implemented purely by motion rules, and patches it with a 'boundary-memory' counter intended to restore unitarity. The central operator, however, is not unitary as written, which breaks the paper's main claim.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Eq. 5 is not unitary: orthogonal inputs |0,0⟩|n−1⟩ and |−1,1⟩|n−2⟩ have overlapping images (inner product 1/√2), and |0,0⟩|0⟩ has no preimage.","rationale":"The central claim is that enforcing unitarity on the motion rule necessitates a boundary-memory and that Eq. 5 realizes it. Everything downstream — 'event-horizon-like' freezing, Page-curve-like entropy, entanglement generation — is presented as a consequence of this unitary evolution. The reader's weakest-assumption identification is correct: Eq. 5 fails the definition of unitarity. I verified the counterexample directly from the displayed sums: for P_same, the term Σ_{x2−2≤x1≤x2+2} P_{x1,x2} ⊗ Σ_{d=0}^{n−2}|d+1⟩⟨d| sends |−1,1⟩|n−2⟩ to |−1,1⟩|n−1⟩, while the final P_same term (1/√2)Σ_{x1=x2}{...} sends |0,0⟩|n−1⟩ to a superposition containing |−1,1⟩|n−1⟩ with amplitude 1/√2. The images of two orthonormal inputs are not orthogonal, so U^†U≠I. Also, no term maps any same-spin state to |0,0⟩|0⟩, so the map is not onto. These are internal inconsistencies of the stated equation, not disagreements with external consensus. The paper's own §2.3 admits the memory completion is non-unique, which weakens the 'necessity' narrative, but the decisive issue is the implemented operator not being unitary. Because the rest of the paper uses this operator to compute entropies, the conclusions do not follow from the claimed premise. A simple analytic recomputation of the inner product settles the objection; no numerical simulation is required.","tokens_in":11365,"tokens_out":7569,"duration_ms":69527,"concrete_test":"Compute U^†U on the two basis states |↑↑,0,0,n−1⟩ and |↑↑,−1,1,n−2⟩ using Eq. 5: evaluate the inner product of their images. If it is 1/√2 (as follows from the P_same third and fourth terms), the operator is not an isometry. Optionally, attempt to solve U|ψ⟩ = |↑↑,0,0,0⟩ on a truncated lattice (|x|≤L, n=3) and verify no solution exists, confirming non-surjectivity. Either check suffices to falsify the unitarity assertion in §2.3.","verdict_should_be":"UNCHANGED","load_bearing_attack":"Equation 5's Û_int is claimed to be unitary ('the operator remains unitary', §2.3), but it is not an isometry on H_c⊗H_p⊗H_b. For the same-spin sector, consider orthogonal inputs |ψ1⟩ = |0,0⟩_p|n−1⟩_b and |ψ2⟩ = |−1,1⟩_p|n−2⟩_b. The fourth term of P_same maps |ψ1⟩ to (1/√2)(|−1,1⟩ + |1,−1⟩)|n−1⟩; the third term maps |ψ2⟩ to |−1,1⟩|n−1⟩. Their images have inner product 1/√2, not 0, so Û_int cannot be unitary (nor even an isometry). Surjectivity also fails: |0,0⟩|0⟩ has no pre-image under any same-spin term. The repulsive branch at memory |n−1⟩ maps separation d to d+2 on the infinite lattice, leaving d=1,2 without preimages. Because the paper's central narrative — that unitarity constraints necessitate boundary-memory, and that the resulting dynamics are unitary — rests entirely on Eq. 5 being a legitimate unitary evolution, this failure invalidates the main construction. The Section 3 entanglement computations may be algebraically correct for this non-unitary map, but they cannot support the claimed 'emergent' unitary dynamics. The manuscript itself concedes the memory completion is non-unique (§2.3), but that concession does not repair the non-unitarity of the implemented operator.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper proposes a minimal discrete-time quantum walk of two particles on a 1D lattice, with spin-dependent attraction (identical spins) and repulsion (opposite spins). The authors argue that imposing unitarity on a naive motion-based interaction rule fails, and that restoring unitarity requires an auxiliary 'boundary-memory' degree of freedom that acts as a counter while the particles are trapped in an 'internal region'. They then study entanglement between the particles and this memory, reporting event-horizon-like freezing and Page-curve-like entanglement entropy. The central technical object is the operator Û_int in Eq. (5), which the paper asserts is unitary.","tokens_in":11759,"tokens_out":3975,"duration_ms":39657,"significance":"If the construction were correct, the paper would provide a simple toy model in which unitarity constraints force an auxiliary memory degree of freedom and give rise to non-trivial entanglement dynamics, possibly relevant to quantum-information and black-hole-inspired toy models. The paper also makes an effort to distinguish its motion-based interaction from earlier phase-based interacting quantum walks, and the entanglement calculations in Section 3 are explicit and reproducible. However, the central premise is invalid: the operator in Eq. (5) is not unitary on the stated Hilbert space. Since the paper's key claims—that unitarity necessarily induces boundary-memory and that the resulting dynamics are unitary—rest entirely on this operator, the main result is not supported. The manuscript text itself concedes that the boundary-memory completion is non-unique and chosen as 'the simplest implementation' (§2.3), further undermining the 'emergent' narrative.","major_comments":[{"comment":"The operator Û_int is not an isometry. Consider orthogonal inputs |ψ1⟩ = |0,0⟩_p |n−1⟩_b and |ψ2⟩ = |−1,1⟩_p |n−2⟩_b. The same-position term in the |n−1⟩⟨n−1| branch maps |ψ1⟩ to (|−1,1⟩ + |1,−1⟩)/√2 |n−1⟩. The internal-region increment term Σ_{d=0}^{n−2}|d+1⟩⟨d| maps |ψ2⟩ to |−1,1⟩|n−1⟩. These images have inner product 1/√2, not 0, so the map cannot be unitary. The statement in §2.3 that 'the operator remains unitary' is therefore false.","section":"§2.3, Eq. (5)"},{"comment":"The map is not surjective. The state |0,0⟩_p |0⟩_b has no preimage: external shift terms cannot reach (0,0), the internal increment term only increases the memory index from d≥0 to d≥1, and the same-position and repulsive branches produce separation-2 or larger states. Thus the image of the full Hilbert space does not contain |0,0⟩|0⟩, so Û_int is not unitary (nor even invertible).","section":"§2.3, Eq. (5)"},{"comment":"The central narrative that unitarity constraints necessarily require boundary-memory is unsupported. The text explicitly states the completion is 'not a unique solution' and is 'the simplest implementation' (§2.3), so the memory is introduced by hand rather than derived. Moreover, because Eq. (5) is non-unitary, the Section 3 entanglement and Page-curve-like results are computed with a non-unitary evolution and cannot support the claimed 'emergent boundary-memory' or 'event-horizon-like' behavior. The paper would need a genuinely unitary operator and a proof (or at least a clear argument) that memory is forced, not merely one possible completion.","section":"§2.3 and §4"}],"minor_comments":[{"comment":"The citation 'Omaret al.' should read 'Omar et al.'; there are a few similar typographical issues in the reference list.","section":"Introduction, §1"},{"comment":"The phrase 'entanglement between subsystem' should be 'between subsystems' for grammatical clarity.","section":"Abstract"},{"comment":"In the sentence 'the boundary-memory does not start counting yet', 'yet' is misleading since the memory will start counting later; consider rephrasing to 'has not started counting'.","section":"§3.2.2"},{"comment":"The operator is dense and hard to parse because multiple sums and projectors are combined without a clear block structure. Even in a revision with a corrected unitary operator, a basis-state action table would greatly improve readability.","section":"Eq. (5)"}],"recommendation":"reject","confidential_remarks":"The reader's report and the stress-test calculation agree on the decisive defect: Eq. (5) is not unitary, and the paper's central claims depend on that operator. The manuscript would need a fundamentally new construction, not just local edits, before it could be considered for publication. I recommend rejection."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Colleague,\n\nThe central claim fails: Eq. 5 is not unitary. The stress-test counterexample is correct. Take same-spin states |0,0⟩|n−1⟩ and |−1,1⟩|n−2⟩; the first maps to (1/√2)(|−1,1⟩+|1,−1⟩)|n−1⟩, the second to |−1,1⟩|n−1⟩, so orthogonal inputs have images with inner product 1/√2. Also |0,0⟩|0⟩ has no preimage, so the map is neither an isometry nor surjective. The assertion 'the operator remains unitary' (Section 2.3) is false as written.\n\nThis matters because everything downstream depends on that premise. The paper's arc—that unitarity constraints necessarily force a boundary-memory degree of freedom, and that the entropy of particles versus memory behaves like a Page curve—is load-bearing on Eq. 5. The Section 3 entanglement calculations are algebraically consistent with the non-unitary linear map, but they do not support the claimed unitary dynamics.\n\nCredit where due: the paper identifies a genuine obstruction in motion-based interacting DTQWs—naive attractive rules are many-to-one, and simple fixes like reflecting or freezing don't restore invertibility. That is a real and clearly explained observation, and the worked states in Section 3 are new for this model. The authors also honestly note that the memory completion is non-unique and 'the simplest implementation' (Section 2.3). The problem is that non-uniqueness doesn't help: the implemented operator is still not unitary.\n\nThe bigger conceptual issue is that the memory is not actually 'emergent' from unitarity; it is inserted by hand, and the necessity claim is asserted rather than proven. A no-go argument would be needed to show that all unitary completions require such a register. The event-horizon and Page-curve analogies are overplayed in the abstract, even though the text later disclaims the black-hole connection.\n\nI wouldn't cite the unitarity results. If the authors repair the operator—for example, by distinguishing the freezing step from the counting step or adjusting the boundary term—the model could become a modest but acceptable contribution. As submitted, the load-bearing math fails and I'd recommend rejection. A kind rejection with the counterexample would give the authors a concrete path to fix it.\n\nFor peer review: this is a borderline case. The paper is not important enough or formally grounded enough to justify referee time over the clear error; I'd desk reject with the counterexample.","headline":"The paper's central construction fails: Eq. 5 is not unitary, so the 'emergent boundary-memory' narrative is unsupported.","tokens_in":12357,"tokens_out":6979,"would_cite":false,"duration_ms":69518,"reading_group":"maybe","serious_thinker":"no","would_accept_peer_review":false},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"This paper claims that enforcing unitary (reversible) evolution on a minimal two-particle attraction rule on a one-dimensional lattice forces an auxiliary 'boundary-memory' counter, whose freezing, particle-memory entanglement, and entropy","keywords":["two-particle quantum walk","unitarity","boundary-memory","event-horizon-like","Page curve","entanglement dynamics","lattice model","quantum cellular automata"],"falsifier":"Apply Eq. 5 to the two orthogonal basis states |↑↑⟩|−1,1⟩|n−2⟩ and |↑↑⟩|0,0⟩|n−1⟩: both images have nonzero overlap along |↑↑⟩|−1,1⟩|n−1⟩ (inner product 1/√2), so the operator is not an isometry; also check whether |x,x⟩|0⟩ has any pre-image. Either failure settles that Eq. 5 is not unitary as written.","tokens_in":11119,"feed_emoji":"⚛️","tokens_out":6624,"duration_ms":68021,"temperature":0.7,"pith_summary":"The paper tries to establish that a deliberately minimal interaction rule — two particles on a one-dimensional lattice that move one site per step, attracting when their spins match and repelling when they differ — cannot be made reversible without an extra degree of freedom. It argues that the naive motion rule is many-to-one when particles come within two lattice sites, and that the cheapest unitary completion is a finite 'boundary-memory' counter that freezes the particles at the boundary of that internal region, counts upward, and ejects them once it saturates. If this construction holds, the same constraint that makes quantum dynamics reversible also manufactures a hidden memory degree of freedom, and its entanglement with the particles reproduces qualitative features of black-hole information dynamics without any gravity. The demonstration is explicit: small superpositions are evolved by hand to show horizon-like freezing, memory-mediated entanglement, and a Page-curve-shaped entropy profile. The entire derivation, however, rests on the assertion that the modified evolution operator in Eq. 5 is exactly unitary.","feed_headline":"Reversibility forces a memory clock into two attracting particles","feed_subtitle":"A minimal lattice model shows reversible dynamics demands boundary-memory; its entropy grows and falls like a Page curve.","key_machinery":"The central object is the operator of Eq. 5, a conditional shift (motion-based) unitary candidate on the combined Hilbert space of spins, positions, and the boundary-memory register. Its key pieces are: (i) in the external region (separation greater than two lattice sites), same-spin particles move one site closer while the memory stays in its ground state; (ii) in the internal region (separation of two sites or less), particle motion halts and the memory increments like a counter; (iii) when the counter is full, same-spin particles are ejected outward, effectively switching attraction to repulsion; (iv) opposite-spin particles repel at all separations. The memory register is what absorbs th","core_discovery":"The central claim, stated on the paper's own terms, is that unitarity is not merely a constraint on an existing model but a generative principle: for spin-parity-dependent attraction and repulsion implemented purely through conditional shifts, reversible evolution is impossible unless the Hilbert space is enlarged. The authors identify the obstruction precisely: when two attracting particles close to within two lattice sites, multiple configurations converge to one, so no unitary operator on position and spin alone exists. The proposed resolution is to append a finite memory register, 'boundary-memory,' whose evolution is a clock: while the particles are frozen in the internal region the mem","pith_inferences":["Editorial inference: the many-to-one obstruction is generic for deterministic shift rules on a line; any local unitary completion that avoids an enlarged Hilbert space would have to abandon strict one-site-shift kinematics, so similar auxiliary registers should appear in other collision-based cellular automaton models, not just this one.","Editorial inference: because the paper's proof that Eq. 5 is unitary is only asserted, the phrase 'restoring unitarity necessarily requires' is not yet established; a corrected operator or a proof of isometry would be needed to make the emergence claim airtight.","Editorial inference: the finite memory size n controls how long particles remain frozen; a direct experiment or simulation varying n should show the entropy peak height growing like log n and the plateau width growing linearly, a signature that would distinguish this mechanism from phase-induced interaction models.","Editorial inference: the Page-curve analogy suggests that any unitary system with temporary information trapping and a 'clock' degree of freedom will show similar entropy dynamics; this could be tested in discrete-time quantum walk implementations on photonic or trapped-ion hardware."],"forward_implications":["If Eq. 5 is unitary, any motion-only two-particle attraction rule on a line needs at least one auxiliary degree of freedom once separations drop to two sites or less; a strictly position-spin Hilbert space is insufficient.","The boundary-memory acts as a physical clock: particles freeze at the event-horizon-like boundary for a finite number of steps, then are ejected, so the model realises temporary information trapping inside a unitary circuit.","Entanglement between the particles and the memory is produced only by superpositions of different entry times; fully localised initial states remain unentangled.","The reduced memory entropy can rise to a maximum and return to zero, producing a Page-curve-shaped profile; with spin superposition the entropy plateaus instead.","Post-selecting on a memory outcome can leave the two particles entangled, so the boundary-memory is a resource for entanglement generation between otherwise non-interacting particles."],"fun_headline_variants":["Unitarity, not forces, creates boundary-memory in two-particle model","Reversible particles need a memory clock to avoid merging","Two-particle lattice: unitarity spawns event-horizon-like behavior","Minimal quantum model: reversibility implies a hidden memory space"],"cache_read_input_tokens":2304,"weakest_assumption_plain":"The load-bearing premise is that the modified evolution operator of Eq. 5 is exactly unitary; the paper asserts this without proof, and as written the map is not an isometry (two orthogonal inputs can share an image), so if the unitarity claim cannot be repaired, the boundary-memory is not shown to be forced by reversible evolution.","fun_headline_variants_meta":{"raw":{"variants":["Unitarity, not forces, creates boundary-memory in two-particle model","Reversible particles need a memory clock to avoid merging","Two-particle lattice: unitarity spawns event-horizon-like behavior","Minimal quantum model: reversibility implies a hidden memory space"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000209,"raw_usage":{"total_tokens":1156,"prompt_tokens":568,"completion_tokens":588,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":312,"completion_tokens_details":{"reasoning_tokens":514}},"tokens_in":312,"tokens_out":588,"duration_ms":7463,"temperature":1.0,"reasoning_tokens":514,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-01T11:30:37.191571+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Apply Eq. 5 to the two orthogonal basis states |↑↑⟩|−1,1⟩|n−2⟩ and |↑↑⟩|0,0⟩|n−1⟩: both images have nonzero overlap along |↑↑⟩|−1,1⟩|n−1⟩ (inner product 1/√2), so the operator is not an isometry; also check whether |x,x⟩|0⟩ has any pre-image. Either failure settles that Eq. 5 is not unitary as written.","supporting_citations":[],"review_version":1}