Hypnopaedia-Aware Machine Unlearning via Psychometrics of Artificial Mental Imagery

arXiv 2410.05284 ↗First posted 29 Sep 2024cs.CR cs.AI cs.CVcited by 0

On the map: LLM safety and security → Backdoor and data poisoning attacks

Abstract (arXiv)

Neural backdoors represent insidious cybersecurity loopholes that render learning machinery vulnerable to unauthorised manipulations, potentially enabling the weaponisation of artificial intelligence with catastrophic consequences. A backdoor attack involves the clandestine infiltration of a trigger during the learning process, metaphorically analogous to hypnopaedia, where ideas are implanted into a subject's subconscious mind under the state of hypnosis or unconsciousness. When activated by a sensory stimulus, the trigger evokes a conditioned reflex that directs a machine to mount a predetermined response. In this study, we propose a cybernetic framework for constant surveillance of backdoor threats, driven by the dynamic nature of untrustworthy data sources. We develop a self-aware unlearning mechanism to autonomously detach a machine's behaviour from the backdoor trigger. Through reverse engineering and statistical inference, we detect deceptive patterns and estimate the likelihood of backdoor infection. We employ model inversion to elicit artificial mental imagery, using stochastic processes to disrupt optimisation pathways and avoid convergent but potentially flawed patterns. This is followed by hypothesis analysis, which estimates the likelihood of each potentially malicious pattern as the true trigger and infers the probability of infection. The primary objective of this study is to maintain a stable state of equilibrium between knowledge fidelity and backdoor vulnerability.

Key facts. No citing papers in the map yet (data through 14 Oct 2026). Placed in Backdoor and data poisoning attacks, one of 12 directions in LLM safety and security. Builds on 55 of its 90 references that are in the map. First posted on arXiv 29 Sep 2024.

Builds on

55 of 90 references.

Closest ancestors · 2

Same direction, newest first

Other references · 53

2015Deep Residual Learning for Image Recognitioncited by 52,096
2015Distilling the Knowledge in a Neural Networkcited by 10,324
2016Communication-Efficient Learning of Deep Networks from Decentralized Datacited by 7,669
2014Explaining and Harnessing Adversarial Examplescited by 7,364
2017Towards Deep Learning Models Resistant to Adversarial Attackscited by 6,415
2013Intriguing properties of neural networkscited by 5,155
2016Overcoming catastrophic forgetting in neural networkscited by 4,787
2019CutMix: Regularization Strategy to Train Strong Classifiers with Localizable Featurescited by 2,884
2016Deep Learning with Differential Privacycited by 2,728
2019Machine Unlearningcited by 1,061
2016Adversarial Machine Learning at Scalecited by 936
2018How To Backdoor Federated Learningcited by 765

38 more

2018Fine-Pruning: Defending Against Backdooring Attacks on Deep Neural Networkscited by 580
2018Spectral Signatures in Backdoor Attackscited by 484
2019STRIP: A Defence Against Trojan Attacks on Deep Neural Networkscited by 459
2018Detecting Backdoor Attacks on Deep Neural Networks by Activation Clusteringcited by 456
2020Backdoor Learning: A Surveycited by 426
2020Invisible Backdoor Attack with Sample-Specific Triggerscited by 353
2021Neural Attention Distillation: Erasing Backdoor Triggers from Deep Neural Networkscited by 324
2020Input-Aware Dynamic Backdoor Attackcited by 313
2019Can You Really Backdoor Federated Learning?cited by 255
2021Adversarial Neuron Pruning Purifies Backdoored Deep Modelscited by 235
2019Detecting AI Trojans Using Meta Neural Analysiscited by 199
2019Adversarial Training and Robustness for Multiple Perturbationscited by 183
2021Adversarial Unlearning of Backdoors via Implicit Hypergradientcited by 166
2020Witches' Brew: Industrial Scale Data Poisoning via Gradient Matchingcited by 163
2017Neural Trojanscited by 153
2020Dynamic Backdoor Attacks Against Machine Learning Modelscited by 151
2019Universal Litmus Patterns: Revealing Backdoor Attacks in CNNscited by 146
2019Online Meta-Learningcited by 137
2018Backdoor Embedding in Convolutional Neural Network Models via Invisible Perturbationcited by 134
2020Backdoor Attacks Against Deep Learning Systems in the Physical Worldcited by 134
2021SPECTRE: Defending Against Backdoor Attacks Using Robust Statisticscited by 130
2019Demon in the Variant: Statistical Analysis of DNNs for Robust Backdoor Contamination Detectioncited by 124
2020RAB: Provable Robustness Against Backdoor Attackscited by 118
2020Deep Partition Aggregation: Provable Defense against General Poisoning Attackscited by 115
2020Certified Robustness to Label-Flipping Attacks via Randomized Smoothingcited by 108
2019Robust Anomaly Detection and Backdoor Attack Detection Via Differential Privacycited by 95
2020Practical Detection of Trojan Neural Networks: Data-Limited and Data-Free Casescited by 95
2020Intrinsic Certified Robustness of Bagging against Data Poisoning Attackscited by 94
2020Strong Data Augmentation Sanitizes Poisoning and Backdoor Attacks Without an Accuracy Tradeoffcited by 89
2020Local and Central Differential Privacy for Robustness and Privacy in Federated Learningcited by 78
2020On Certifying Robustness against Backdoor Attacks via Randomized Smoothingcited by 77
2020DeepSweep: An Evaluation Framework for Mitigating DNN Backdoor Attacks using Data Augmentationcited by 77
2020Certified Robustness of Nearest Neighbors against Data Poisoning and Backdoor Attackscited by 60
2019Model Agnostic Defence against Backdoor Attacks in Machine Learningcited by 52
2021Poison Ink: Robust and Invisible Backdoor Attackcited by 50
2023UNICORN: A Unified Backdoor Trigger Inversion Frameworkcited by 46
2020One-pixel Signature: Characterizing CNN Models for Backdoor Detectioncited by 34
2021Topological Detection of Trojaned Neural Networkscited by 34

50 most cited of 53 All 53 (JSON)

Built on by

Not cited yet.

Similar, not cited

Close in wording, no citation either way