PaperlayerView exact revision on arXiv

Source-linked AI summary

Multi-Jurisdictional Legal Identity Assurance for Capability Gating: A Design-Science Proposal for Tiered, Reusable Identity Assurance of Natural, Juridical, and Machine Entities

Walter Kurz

arXiv:2610.00287v1q-fin.GNcs.AIcs.CRcs.CY

TL;DR

Flat maximum verification and fragmented assurance frameworks do not provide the act-proportionate, reusable assurance state described in the paper’s motivation. The paper proposes a typed, jurisdiction-aware model that separates assurance from capability gates and records reliance for later evaluation. Its comparative evaluation finds the proposed grid’s primitives map directly to the six stated requirements, unlike the evaluated alternatives.

  • Problem

    The paper identifies a missing mapping from requested capabilities or dispute paths to the minimum evidence needed for compliance, proportionality, and recourse.

    Source: §4.9 · p. 23

  • Method

    The paper proposes a reusable assurance artifact with typed entities, a two-axis grid, time-indexed jurisdiction, and liability-allocated reliance records.

    Source: §6 · p. 25

  • Results

    The proposed grid is the evaluated design whose primitives map directly to proportional demand, capability separation, entity imputation, granular disclosure, jurisdictional continuity, and allocated reliance.

    Source: §6 · p. 25

  • Takeaways & Limitations

    The evaluation argues that preserving cell and event structure keeps distinctions addressable in the architecture rather than shifting them into policy prose, manual exceptions, or institutional trust.

    Source: §5.1 · p. 24

    The paper specifies the model but provides no reference implementation, operational governance process, or empirical pilot.

    Source: §7 · p. 26

Abstract

from arXiv · show

Identity assurance is the cost a digital system pays for dishonesty and uncertainty: it exists to make acts attributable when not everyone can be trusted at their word. A common way to pay that cost is flat maximum verification, asking each participant to meet a single high level of identification at entry, before any capability is exercised. Paid on everyone, it over-collects, excludes participants who cannot meet a bar they never needed to clear, taxes every interaction with the cost of the rarest high-risk case, and binds the strength of identification to the activity it unlocks. Existing frameworks compound this by fixing a small number of per-credential levels inside a single legal space and binding each verification to the institution that performed it, leaving cross-border reuse and the tension between data erasure and evidentiary retention unaddressed. This paper develops, as a design-science proposal, a tiered and reusable model of identity assurance for natural, juridical, and machine entities across jurisdictions. Reading the problem through systems theory, where a system changes only when an entity acts, the model holds the assurance state apart from the capability gate that consumes it, so that identity demand follows the act and the weight of its consequences rather than mere presence: a participant may take part with minimal disclosure and supply more only as an act requires. It comprises a typed entity taxonomy, a two-axis coordinate of disclosed assertion scope and source of information, jurisdiction as a time-indexed attribute of the entity, and reliance recorded as bitemporal, liability-allocated, point-in-time snapshots. Requirements are derived from anti-money-laundering, electronic-identity, and data-protection law, and the proposal is evaluated against flat maximum verification, per-credential level-of-assurance designs, and institutional reusable-KYC reliance.

1 Introduction

The introduction argues that identity assurance should attach to consequential acts rather than enrolment, and that a reusable, jurisdiction-aware assurance state must remain distinct from the capability it gates. It proposes this layer for natural, juridical, and machine entities and evaluates it against existing approaches.

1 Introduction

  • Flat maximum verification over-collects, excludes participants, adds friction, and couples identification strength to the capability granted.It also stores a terminal verified/not-verified outcome that cannot later be queried, replayed, or selectively disclosed.Source: §1 · p. 2
  • Existing assurance frameworks provide levels but remain per-scheme, per-credential, or per-service, while leaving jurisdictional change over time unrepresented.A person who moves may have to abandon an account and re-enrol or keep using it under a potentially inapplicable regime.Source: §1 · p. 2
  • Identity demand belongs to the act and the weight of its consequences, not to mere enrolment, because dormant entities create no legally relevant event.The introduction frames this claim through systems theory, in which a system changes when an entity acts.Source: §1 · p. 2
  • Digital communication separates a sender’s identifier from its source, making identity assurance an explicit reconstruction of what physical co-presence once supplied.The same fact confirmed remotely and in physical presence counts as different evidence classes.Source: §1 · pp. 2–3
  • The model includes machine entities while anchoring accountability for their acts in a natural or juridical entity capable of bearing consequences.Source: §1 · p. 3
  • The proposal separates assurance state from capability permission and specifies a reusable, jurisdiction-aware layer evaluated against three alternatives.Its design records each reliance as a point-in-time, liability-allocated snapshot that can be audited under the law and policy version in force at the time.Source: §1 · p. 3

2 Related Work

Related work supplies assurance levels, wallets, organisational identifiers, institutional KYC reuse, and holder-presented credentials, but does not combine them into a portable legal-assurance layer spanning entity types and jurisdictions.

2 Related Work

  • eIDAS and the European Digital Identity Framework provide scheme-level assurance infrastructure, while NIST separates proofing, authentication, and federation; none supplies the proposed portable, jurisdiction-aware state across services.NIST federation reuses assertions inside a pre-agreed trust framework rather than carrying a holder-presented assurance state across legal contexts.Source: §2 · pp. 3–4
  • The LEI identifies legal persons and the vLEI binds officer roles, but neither bounds an entity’s assurance by accountable persons’ assurance and valid delegation.The proposed inheritance bound supplies that missing constraint.Source: §2 · p. 4
  • Institutional reusable KYC exchanges due-diligence materials but keeps reliance institution-bound and responsibility with the relying institution.The proposal instead treats the reusable object as a holder-carried assurance history with allocated, point-in-time reliance records.Source: §2 · p. 4
  • Verifiable-credential standards and self-sovereign-identity work address credential presentation and selective disclosure, while this paper adds a legal-assurance layer around them.Source: §2 · p. 4
  • The missing object joins progressive assurance, jurisdiction awareness, capability separation, typed entities, bounded juridical assurance, and replayable reliance.Source: §2 · p. 4

3 Contribution

The contribution is a construct and an ontology-package artifact that separates assurance from capability, represents evidence on two axes, and records reusable reliance across jurisdictions. The proposed architecture is offered for evaluation, not as a fixed choice of levels or tiers.

3 Contribution

  • The construct treats the entity—not a customer, credential holder, or account—as the assurance subject, and distinguishes natural, juridical, and machine entities.For juridical entities, the inheritance bound caps assurance by accountable natural persons’ assurance and valid delegation.Source: §3 · p. 4
  • The artifact keeps reusable assurance state apart from capability gates, allowing one state to pass a low-disclosure gate, fail a higher one, and be replayed for audit.Source: §3 · p. 5
  • The assurance coordinate separates disclosed assertion scope from the source standing behind each assertion, so disclosure and evidentiary strength can be graded independently.The number of levels on either axis is a design parameter, not a claimed contribution.Source: §3 · p. 5
  • Jurisdiction is a time-indexed entity attribute, and each reliance event records claim state, valid and record times, governing law, requirement version, parties, and liability boundary.Source: §3 · p. 5
  • The paper claims the architecture and its separations, not any particular setting of its levels or tiers.Source: §3 · p. 5

4 Research

The research frames identity assurance as infrastructure for imputing consequential acts to accountable points. It derives proportionate assurance demands from the act, separates assurance from permission, and specifies reusable evidence for capability gates.

4.1 Identification thresholds for capability-bearing interactions

  • The design supports low-consequence participation with minimal identity payload and lets higher-consequence gates require stronger or fresher evidence or explicit regulatory mapping.Verified claims can be presented again with consent, source, time, and scope, shifting repeated collection toward auditable reliance.Source: §4.1 · p. 5
  • Demanded assurance rises with an act’s exposure to deception and consequence weight, from an identifier floor to maximum verification at the limiting pole.Flat maximum verification fixes every act at that maximum; the proposed design follows the act-specific demand instead.Source: §4.1 · p. 6
  • Identity assurance makes consequential acts imputable to an accountable point; the system registers acts and conditions what may follow rather than knowing a customer.The paper treats natural and juridical persons as kinds of legal point of imputation.Source: §4.1 · p. 6
  • A system demands identity only to the degree needed to impute an action’s consequences to an accountable point.Source: §4.1 · p. 6
  • Assurance strength and permission are independent: the same assurance state may pass one capability gate, fail another, and later support an audit.Source: §4.1 · p. 6
  • The assurance coordinate separates which facts an act requires from how strongly each fact is confirmed, using assertion and source-of-information axes.An act’s demand is a set of assertion classes and minimum confirmation strengths, rather than one shared level.Source: §4.1 · pp. 6–7

4.2 Entity taxonomy for assurance evaluation

The taxonomy classifies assurance subjects by suability and source of behaviour, then treats each type as a distinct imputation regime. It links type-specific evidence and accountability to the act being assessed.

4.2 Entity taxonomy for assurance evaluation

  • An entity bears an assurance state; it becomes an actor when an act occurs and consequence is imputed to it.Technical action and legal identity may differ, particularly for machines whose accountability runs through surrounding natural or juridical persons.Source: §4.2 · pp. 7–8
  • Two cuts—suability and source of behaviour—yield natural, juridical, and machine entities; the remaining branch is an inert object outside the model.The figure presents the resulting assurance grid as a map, not a ladder of ranks.Source: §4.2 · pp. 7–8, Figure 1 · p. 8
  • Natural and machine entities originate behaviour, while juridical entities act only through natural persons or mandated machines.Malice originates in a behaviour-bearing entity; juridical fault is an imputed state rather than an original behaviour source.Source: §4.2 · p. 8
  • An active juridical entity must have a traceable ownership, control, organ, trustee, or representative path to an identified natural person or fiduciary officeholder.The model keeps an entity without such an accountable endpoint dormant; intermediate links may be juridical entities.Source: §4.2 · p. 9
  • A juridical entity’s assurance cannot exceed that of its accountable natural persons, assessed over the delegation actually used for each act.Jointly required signers are bounded by the weakest; where any one of several may act, the entity may put forward its strongest admissible constellation.Source: §4.2 · p. 9
  • Juridical forms share a catalogue and imputation regime, so their differences are represented as gate-readable predicates rather than separate entity types.Partnership liability and trust or other legal-arrangement treatment are handled through predicates and the relevant fiduciary or trustee.Source: §4.2 · pp. 9–10
  • Machines are described by capability, a versioned trust vector, and control; familiar labels such as bots and smart contracts are not gate inputs.The model retains legal accountability around the machine and admits consequential operation through specified accountable paths.Source: §4.2 · p. 10
  • Disclosure follows entity type: natural persons retain data-protection interests, juridical entities follow their disclosure regime, and machines disclose their attachment path.A worked trace replays the machine’s point-in-time attachment and the applicable law to identify where consequences are directed.Source: §4.2 · p. 11

4.3 Jurisdictional anchoring and temporal attribution

Jurisdiction is a time-indexed attribute of an entity’s assurance state, selected in light of the act being evaluated. Bitemporal records preserve which artifacts and legal frame applied when the act occurred.

4.3 Jurisdictional anchoring and temporal attribution

  • The same entity can retain one assurance state while jurisdictional artifacts change over time; the act selects the relevant anchor and legal predicates.Other jurisdictional artifacts may remain in the state as corroborating, historical, or capability-specific evidence.Source: §4.3 · p. 11
  • Stored facts remain distinct from jurisdiction-relative predicates, which are evaluated under the active anchor and law governing the requested act.The same fact may satisfy one jurisdiction’s gate and fail another’s without changing the fact itself.Source: §4.3 · p. 11
  • Bitemporal records preserve valid time alongside record time so later audits can replay the artifacts, anchor, and predicates applicable to an earlier act.This separates the legal and policy state relevant to an act from the later state in which its record is audited.Source: §4.3 · p. 12

4.4 Requirements for multi-jurisdictional legal identity assurance

The requirements translate the model’s legal and standards pressures into constraints for comparative evaluation. Assertion granularity makes data minimisation structurally enforceable by letting gates request only the identity content they need.

4.4 Requirements for multi-jurisdictional legal identity assurance

  • The requirements are six testable constraints linking legal or standards pressure to the construct that discharges each constraint.The evaluation uses them to assess candidate assurance architectures; AML due diligence supplies an upper mapping for capabilities within its domain.Source: §4.4 · p. 12, Table 1 · p. 13

4.5 Assertion granularity and data minimisation

  • An assertion class is one kind of identity content in a type’s ordered catalogue, separately addressable so an act can request only its required facts.The axis is built from the demand side: each legally distinguishable demand must be separately addressable.Source: §4.5 · p. 12
  • Separate assertion rows make data minimisation enforceable by preventing a gate that needs one item from receiving a packed identity record.Unrequested classes remain structurally undisclosed rather than depending on a policy choice to withhold them.Source: §4.5 · p. 12
  • Only the existence-status row is mandatory; voluntary rows may be empty, and each class is either single-slot or an open set of instances.Functional data stays service-local unless a gate predicate or legal mapping can demand it.Source: §4.5 · p. 12
  • Multiple artifacts in one assertion class increase corroboration rather than tier, with confirmations sharing a failure mode counted together.Three passports remain instances of the government-identity-document class; their issuers may provide independent corroboration.Source: §4.5 · p. 14

4.6 Source-of-information scale

The source-of-information axis records the accountable strength behind each disclosed assertion, alongside its binding to the subject. The shared scale supports three entity-specific assertion catalogues, with unavailable source classes treated as inapplicable rather than failed.

4.6 Source-of-information scale

  • S1–S7 order confirmation classes by accountable weight, with co-presence outranking remote confirmation within tiers and recognised public or qualified sources above trained human verifiers.The confirmation’s value at a gate also depends on its binding and freshness, which the relying party derives rather than stores as a score.Source: §4.6 · pp. 14–15
  • A confirmation event fills only the row-facts it examined; cells filled together are correlated, and confirmation chains are bounded by their weakest required link.If an event is spoofed, its covered cells retract together.Source: §4.6 · p. 15
  • When a source class cannot exist for an item, its cell is inapplicable rather than failed, allowing assurance through the strongest available recognised channel.The rule avoids penalising entities for lacking a source they could never use.Source: §4.6 · p. 16

4.7 Per-type assertion catalogues

  • The natural catalogue orders N0–N10 by willingness to reveal, from mandatory existence status through increasingly sensitive assertions to voice and face patterns.Biometric patterns are last because disclosure is costly and least reversible.Source: §4.7 · p. 16
  • The juridical catalogue assigns J0–J7 to status, legal name and form, founding basis, register, authority, capacity, and accountable persons.A gate combines the entity’s cells with relevant natural persons’ grids and the delegation or office relation connecting them.Source: §4.7 · pp. 16–17
  • The machine catalogue uses M0 for existence and accountability attachment, while M1–M3 describe capability and trust and control remain cell descriptors.M3 is reserved for goal-setting beyond a human-authored mandate.Source: §4.7 · p. 17

4.8 Legal identity-assurance coordinate and event model

The assurance coordinate combines disclosed assertion scope with the source strength behind each assertion, preserving a set of addressable cells rather than a single score. Confirmation and reliance events retain provenance and decision context so that evidence can be reused and later decisions replayed.

4.8 Legal identity-assurance coordinate and event model

  • The first axis records which identity assertions are available, legally required, provided, and disclosed for a capability context; holding an artifact does not itself disclose it.The second axis records the evidentiary strength of the source behind each disclosed assertion, not the actor’s trustworthiness.Source: §4.8 · p. 18
  • An assurance state is a set of filled assertion-source cells, not a summed score, because a strong cell cannot substitute for a missing one.Each cell retains its confirmation dates and resolves to the artifacts and confirmations behind it.Source: §4.8 · p. 18
  • Folding the grid into a single signal or ladder deletes cell addressability and therefore removes gate demands the full coordinate could express.The paper evaluates these simplified designs as folds of the grid.Source: §4.8 · pp. 18–19, Figure 3 · p. 19
  • Confirmation events preserve source, binding, timing, status, policy, and proof context; cells filled by one event are correlated and do not count as independent confirmations.If the event is revoked or invalidated, its supported cells lose that event as evidence.Source: §4.8 · p. 19
  • Reliance events record the decision, gate, act, applicable law, and liability boundary, with valid time distinguished from record time for later replay.The record is designed to preserve decision context while the relying party receives only the presentation its gate declared.Source: §4.8 · pp. 20–21

4.9 Capability-gate grammar and assurance separability

The section separates reusable identity-assurance state from capability-specific decisions, so assurance demand follows the act and the evidence its consequences require. Gates evaluate both assurance evidence and the participating actor constellation, while disclosure and reliance remain scoped to the requested operation.

Assurance separability

  • Reusable assurance state is distinct from gate decisions, allowing the same actor state to pass a low-disclosure gate, fail a banking gate, and later support audit replay.Gate predicates evaluate the state under the law and policy version in force at the time.Source: §4.9 · p. 21

Gate design

  • Gate predicates attach to legally relevant entities, mandates, credentials, transactions, and conduct, while absent non-required data is neutral rather than a suspicion marker.Adverse decisions should expose their reason category, evidence class, applicable rule, and recourse path.Source: §4.9 · p. 21

Actor constellations

  • Gates evaluate actor constellations, including natural actors, juridical actors acting through natural anchors, and machine participation; standalone machine action is inadmissible in this artifact.The grammar defines six patterns, with some admissibility service-dependent or legally restricted.Source: §4.9 · p. 21

Gate grammar

  • A gate bundles demanded assurance cells and freshness limits, admissible constellations, and governing law; failure affects only the requested capability, not the entity’s standing.The displayed predicate requires an admitted constellation and sufficiently strong, recent evidence for each demanded cell.Source: §4.9 · p. 21

Freshness

  • Validity and eligibility are distinct: genuine, unexpired evidence may still fail a gate’s freshness window, which the relying party evaluates rather than storing as a cell flag.The passage contrasts an eight-month-old notarisation with a six-month account-opening window and a date-of-birth fact with no declared expiry window.Source: §4.9 · p. 22

Disclosure

  • A gate receives only its justified signed evidence package and included lower assertions; a holder may separately signal a stronger assurance ceiling without disclosing higher-tier evidence.Selective-disclosure techniques can support this split, subject to the privacy and correlation limits of the proof format.Source: §4.9 · p. 22

Regulatory mapping

  • Regulatory mappings translate sectoral demands, including AML due diligence, into gate predicates while preserving a general assurance state reusable across other capability requirements.The proposed layer starts from the capability, actor constellation, and evidence needed to make an operation imputable.Source: §4.9 · p. 22

Evidence and mapping

  • Source-diverse evidence supports attribution resilience rather than a general trust score, while capability mappings connect requested acts or disputes to minimum evidence packages.The passages contrast weak social-media traces with stronger legal-identity evidence and describe triangulation as making the submitted state harder to fabricate and easier to cross-check.Source: §4.9 · pp. 22–23

4.10 Comparative design evaluation

The section evaluates four architectures analytically against six requirements, assessing what each can express natively rather than implementation performance. The proposed grid is the only evaluated design whose primitives correspond to all six requirements, while the alternatives retain useful but incomplete capabilities.

Evaluation method

  • The comparison is an analytical assessment of design capacity, not implementation performance, and scores native representation, partial representation, or reliance on external policy or process.The comparator set includes flat maximum verification, per-credential LoA, institutional reusable-KYC reliance, and the proposed grid.Source: §4.10 · p. 23

Scenario readings

  • Scenario readings show the grid expressing different evidence demands over shared state, gate-specific outcomes, accountable machine action, minimal age disclosure, jurisdictional change, and recorded reliance.The scenarios span six requirements, including replaying the prior jurisdictional anchor and recording the relying service, package, rule version, and liability boundary.Source: §4.10 · pp. 23–24

Comparative evaluation

  • The evaluation finds that only the proposed grid has primitives corresponding one-to-one to all six requirements; the other designs lose distinctions involving addressability, event timing, or assurance versus permission.Institutional reusable-KYC reliance supports part of reuse and some juridical due diligence, but remains a sectoral file rather than a portable legal identity-assurance state.Source: §4.10 · p. 24

5 Discussion

The discussion argues that the grid preserves distinctions that one-signal designs shift into policy, manual exceptions, or institutional trust. It then describes how published gate profiles, confirmation events, and shared standards could make those distinctions usable across services and institutions.

5.1 Design rationale and trade-off structure

  • One-signal designs move distinctions out of the architecture and into policy prose, manual exception handling, or institutional trust; the grid keeps them addressable in state, gates, and reliance records.The discussion frames this as the rationale for a grid rather than a ladder.Source: §5.1 · p. 24, Table 9 · p. 24
  • Assurance can accumulate confirmations while gates change with law, risk, policy, or capability; reuse is cumulative for the holder but selective for each relying party.Each new reliance is recorded, distinguishing reusable confirmation from the relying party’s own decision and legal boundary.Source: §5.1 · p. 24

5.2 Institutional implications and adoption conditions

  • A service can publish its gate profile, making data demands, evidence freshness, source strength, and liability boundaries comparable for the same capability.The discussion identifies that comparability as the institutional form of proportionality.Source: §5.2 · pp. 24–25
  • Confirmers’ familiar acts would produce confirmation events bound to explicit rows, source class, valid and record times, policy version, status channel, and proof material.The passage says verifiable-credential standards can carry much of the substrate, with legal identity assurance supplying its semantics.Source: §5.2 · p. 25
  • Regulators could compare gate demands and examine whether they are justified, whether lower-disclosure paths exist where lawful, and whether decisions can be replayed under the applicable rule version.The same grammar covers child safety, AML, corporate signing, consumer refunds, and machine delegation without making their predicates identical.Source: §5.2 · p. 25
  • Adoption depends on network formation and institutional agreement on shared identifiers, source mappings, status channels, proof formats, and role-equivalence tables across jurisdictions.Machine actors also require operators to expose attachment and mandate evidence before automated capability can be accepted as legally meaningful.Source: §5.2 · p. 25

6 Conclusion

The conclusion presents a reusable legal identity-assurance artifact for digital acts and reports an analytical comparison against three alternatives. Its architectural claim is the separation of assurance state from capability gates while retaining cell and event structure.

Conclusion

  • The artifact separates entity, actor, and legal identity, represents assurance through assertion rows and source classes, and treats jurisdiction as time-indexed.Confirmation and reliance are recorded as events with provenance, policy version, parties, timing, and liability boundary.Source: §6 · p. 25
  • The comparative evaluation finds that each alternative folds away at least one distinction needed by the six requirements, while the proposed grid’s primitives map directly to them.The requirements cover proportional demand, assurance-capability separation, N/J/M imputation, granular disclosure, jurisdictional continuity, and allocated reliance.Source: §6 · p. 25
  • The tier count remains a design parameter; the conclusion identifies separation of assurance state from gates and preservation of cell and event structure as the architectural claim.Those structures make the separation usable.Source: §6 · p. 25

7 Limitations and Future Research

The paper identifies governance, strategic-behavior, cost, and scope constraints that must be addressed before deployment. Future work includes testing gate profiles and implementing and evaluating the model.

7 Limitations and Future Research

  • Stable assertion identifiers, source mappings, status channels, proof formats, and role-equivalence tables require standardisation; S6/S7 authority roles need jurisdiction-specific mapping for production comparability.The quality of these governance tasks will determine whether the architecture remains interoperable or fragments into local dialects.Source: §7 · p. 25
  • Actors could game assurance coordinates by learning capability thresholds and manufacturing weak confirmations, so parameter setting requires governance and risk analysis.The model addresses gaming structurally through source classes, binding, confirmation-event correlation, freshness windows, and a confirmation index; a reference implementation should test over-requesting and source shopping.Source: §7 · p. 25
  • The model avoids cheap confirmed-age shortcuts, so gates requiring legally reliable age need sufficiently strong confirmation even when self-declaration is easier.Similar costs arise for high-value jurisdictional predicates, beneficial-ownership paths, machine mandates, and notarial or public confirmations; deployments must balance cost, friction, and risk when choosing lower gates.Source: §7 · pp. 25–26
  • The proposal specifies assurance coordinates, actor grammar, gate semantics, and reliance events, but provides no reference implementation, operational governance process, or empirical pilot.Aggregation, consensus, and mandate mechanics are left to companion papers; future work should implement and test the model, assess selective-disclosure privacy and correlation risks, and measure institutional handling of confirmation events.Source: §7 · p. 26

Declarations

The article provides its publication and availability information, reports editorial-only use of generative AI, and states its funding, conflict, and authorship declarations.

Declarations

  • The 2026 research article is by Kurz in Swissi AI Journal and has DOI 10.5281/zenodo.21901241.The citation lists volume 2026 and article SAIJ-5kdnql4rsq27.Source: Declarations · p. 29
  • The article's data and code are available from the corresponding author on reasonable request.Source: Declarations · p. 29
  • Generative AI was used for editorial work only, while the author verified the content and remains responsible; the research received no external funding and the author declares no conflicts of interest.The author is solely responsible for the article.Source: Declarations · p. 29
  • The article was submitted 05/26, revised 06/26, and published 07/26 under a CC BY 4.0 licence.Source: Declarations · p. 29
Loading 2610.00287v1…