Deterministic Peel Chains for Regulatory Compliance in Autonomous, Offline Environments
- 1Independent researcher, Geneva, Switzerland
Abstract
Financial-crime and regulatory-compliance analysis is increasingly performed with cloud-hosted, machine-learning tooling whose outputs are probabilistic, whose training data and model weights are opaque, and whose operation requires transmitting sensitive subject data to third-party infrastructure. For work that must be reproducible under evidentiary scrutiny and confidential by construction, these properties are liabilities rather than features. This paper describes an alternative architecture, realised in a working reference system (the Forensic & Regulatory Intelligence Suite, FRIS), organised around three commitments: (1) an offline, dependency-free, single-file client architecture, autonomous in the sense that it needs no server, external service or network connection, in which analysis executes entirely on the analyst's own machine and no subject data is transmitted anywhere; (2) a deterministic peel-chain method for tracing layered value transfers, in which value is represented as exact integers, every hop conserves value under an explicit, checked invariant, and the entire trace is byte-for-byte reproducible from its inputs — in deliberate contrast to probabilistic clustering and learned classifiers; and (3) a provenance-first data model in which every reported value is labelled by source, real and synthetic inputs are never mixed silently, and unresolved gaps (unknown ownership percentages, nominee interposition, circular holdings) are surfaced explicitly rather than imputed. We situate the design against the blockchain-analytics, RegTech/SupTech, deterministic-forensics and air-gapped-tooling literatures; specify the peel-chain algorithm and its value-conservation invariant with pseudocode; describe the 23-module engine framework that applies a common deterministic discipline across beneficial-ownership tracing, sanctions and adverse-media screening, an insolvency-recovery waterfall, and a structured compliance knowledge base; and discuss reproducibility, auditability, limitations and future work. No performance benchmarks are claimed, and any numeric example is labelled illustrative.
Preprint; not yet peer reviewed.
Related
- Neuro-Symbolic Regulatory Retrieval and Compliance Review for Railway Construction Plans Adjacent to Operating Lines
- Integrated Predictive and Remote-Sensing Framework for Early Warning and Regulatory Compliance in Environmentally Sensitive Urban Zones
- Regulatory Debrief and NPDES Compliance Analysis for Stormwater Discharge Permit COR042008
- Beyond the Binary of Resistance and Compliance: Systemic Capital, Non-Reactivity, and the Destruction of Coercive Utility in Institutional Environments
- Deterministic Execution Fencing and Write-Ahead Containment for Autonomous Systems
BibTeX
@misc{farmer2026deterministic,
title = {{Deterministic Peel Chains for Regulatory Compliance in Autonomous, Offline Environments}},
author = {Christopher I. V. Farmer},
howpublished = {Zenodo},
year = {2026},
month = oct,
doi = {10.5281/zenodo.23462154},
url = {https://zenodo.org/records/23462154},
note = {CC BY-NC 4.0},
}
RIS
TY - UNPB
TI - Deterministic Peel Chains for Regulatory Compliance in Autonomous, Offline Environments
AU - Christopher I. V. Farmer
PY - 2026
DA - 2026/10/19
PB - Zenodo
DO - 10.5281/zenodo.23462154
UR - https://zenodo.org/records/23462154
AB - Financial-crime and regulatory-compliance analysis is increasingly performed with cloud-hosted, machine-learning tooling whose outputs are probabilistic, whose training data and model weights are opaque, and whose operation requires transmitting sensitive subject data to third-party infrastructure. For work that must be reproducible under evidentiary scrutiny and confidential by construction, these properties are liabilities rather than features. This paper describes an alternative architecture, realised in a working reference system (the Forensic & Regulatory Intelligence Suite, FRIS), organised around three commitments: (1) an offline, dependency-free, single-file client architecture, autonomous in the sense that it needs no server, external service or network connection, in which analysis executes entirely on the analyst's own machine and no subject data is transmitted anywhere; (2) a deterministic peel-chain method for tracing layered value transfers, in which value is represented as exact integers, every hop conserves value under an explicit, checked invariant, and the entire trace is byte-for-byte reproducible from its inputs — in deliberate contrast to probabilistic clustering and learned classifiers; and (3) a provenance-first data model in which every reported value is labelled by source, real and synthetic inputs are never mixed silently, and unresolved gaps (unknown ownership percentages, nominee interposition, circular holdings) are surfaced explicitly rather than imputed. We situate the design against the blockchain-analytics, RegTech/SupTech, deterministic-forensics and air-gapped-tooling literatures; specify the peel-chain algorithm and its value-conservation invariant with pseudocode; describe the 23-module engine framework that applies a common deterministic discipline across beneficial-ownership tracing, sanctions and adverse-media screening, an insolvency-recovery waterfall, and a structured compliance knowledge base; and discuss reproducibility, auditability, limitations and future work. No performance benchmarks are claimed, and any numeric example is labelled illustrative. Preprint; not yet peer reviewed.
ER -