Architectural Backdoors in Deep Learning: A Survey of Vulnerabilities, Detection, and Defense
Fuente:
arXiv
Salvato in:
| Autori principali: | , , |
|---|---|
| Natura: | Preprint |
| Pubblicazione: |
2025
|
| Soggetti: | |
| Accesso online: | |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866908454439878656 |
|---|---|
| author | Childress, Victoria Collyer, Josh Knapp, Jodie |
| author_facet | Childress, Victoria Collyer, Josh Knapp, Jodie |
| contents | Architectural backdoors pose an under-examined but critical threat to deep neural networks, embedding malicious logic directly into a model's computational graph. Unlike traditional data poisoning or parameter manipulation, architectural backdoors evade standard mitigation techniques and persist even after clean retraining. This survey systematically consolidates research on architectural backdoors, spanning compiler-level manipulations, tainted AutoML pipelines, and supply-chain vulnerabilities. We assess emerging detection and defense strategies, including static graph inspection, dynamic fuzzing, and partial formal verification, and highlight their limitations against distributed or stealth triggers. Despite recent progress, scalable and practical defenses remain elusive. We conclude by outlining open challenges and proposing directions for strengthening supply-chain security, cryptographic model attestations, and next-generation benchmarks. This survey aims to guide future research toward comprehensive defenses against structural backdoor threats in deep learning systems. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2507_12919 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Architectural Backdoors in Deep Learning: A Survey of Vulnerabilities, Detection, and Defense Childress, Victoria Collyer, Josh Knapp, Jodie Cryptography and Security Architectural backdoors pose an under-examined but critical threat to deep neural networks, embedding malicious logic directly into a model's computational graph. Unlike traditional data poisoning or parameter manipulation, architectural backdoors evade standard mitigation techniques and persist even after clean retraining. This survey systematically consolidates research on architectural backdoors, spanning compiler-level manipulations, tainted AutoML pipelines, and supply-chain vulnerabilities. We assess emerging detection and defense strategies, including static graph inspection, dynamic fuzzing, and partial formal verification, and highlight their limitations against distributed or stealth triggers. Despite recent progress, scalable and practical defenses remain elusive. We conclude by outlining open challenges and proposing directions for strengthening supply-chain security, cryptographic model attestations, and next-generation benchmarks. This survey aims to guide future research toward comprehensive defenses against structural backdoor threats in deep learning systems. |
| title | Architectural Backdoors in Deep Learning: A Survey of Vulnerabilities, Detection, and Defense |
| topic | Cryptography and Security |
| url | https://arxiv.org/abs/2507.12919 |