P014 Design-flow and fab-process integrity
What detection rate do automated trojan detectors reach on design-flow and fab-process data for unseen trojan families at accelerator scale?
Trojan detection in design flow and fab process data.
Status
Analyzed A desk analysis of the fab-process clause: repurposing fabs' existing process-data anomaly detection and virtual metrology for security, with its stated limits (catches only attacks that show as statistical deviations; high false-positive risk in a noisy fab). No detection rate anywhere. The design-flow clause rests on graph-neural-network detectors trained on netlists reported second-hand (Thorat et al. 2025, not held), which flag patterns resembling known Trojans rather than unseen families; no held document tests any detector at accelerator scale or on unseen trojan families. Logic locking and other obfuscation are described in the same paper as broken by SAT-style attacks, but those attack papers are not held.
A more scalable approach is to integrate verification into the fabrication process itself. Modern fabs already collect enormous amounts of process data for yield and quality control, including sensor readings, optical inspections of masks and wafers, and electrical test parameters. By applying statistical anomaly detection to this data, one can attempt to catch deviations that might indicate tampering.
Posed by
- Ilhan, Withers, Gietz & Harack (2026), Verifiable Semiconductor Manufacturing (Oxford AIGI):
they do not address compromises introduced earlier in the design flow, such as malicious RTL, third-party IP, or compromised EDA tools.
(§4.1.2, p.12, L615–617) (source states the problem but not the quantity the question asks for)
Sub-questions
| ID | Sub-question | What the sources report |
|---|---|---|
| P014.1 | What detection rate do GNN trojan detectors reach on unseen trojan families at accelerator-scale designs? | partly answered |
| P014.2 | Is there a logic-locking scheme with no published break after a stated number of years, and can a locked netlist be verified (formal equivalence) without the key or interfering with post-hoc verification? | partly answered |
| P014.3 | In seeded-tampering studies on real fab process data, what TPR at fixed FPR does inline / virtual-metrology anomaly detection achieve, including trojans kept within control limits? | implied by a source, not stated as open |
| P014.4 | Are design data in transit and conventional fab cyber risks already adequately defended? | a source suggests an answer (not checked) |
Possibly relevant work found by automated search
Found by AI agents searching for this problem's sub-questions, mostly by keyword and citation in the OpenAlex scholarly database and on the web; a second AI pass reviewed each one, and we spot-checked some by hand. They seemed reasonably related, as far as one can tell from outside each paper's field. Some come from neighbouring fields, such as arms control and financial audit, and bear on the problem by analogy. Not reviewed against the status above; a paper listed here does not change that status. 928 of 1052 quotes checked word for word against the paper's text.
For P014.1: What detection rate do GNN trojan detectors reach on unseen trojan families at accelerator-scale designs?
For P014.4: Are design data in transit and conventional fab cyber risks already adequately defended?
Data
- problems.csv — one row per problem
- problem_status.csv — status, scope and evidence per problem
- problem_sources.csv — the posed-by quote and other documents per problem
- leaves.csv — all sub-questions, including those without a checkable form
- leaf_sources.csv — every source row behind each sub-question