ImpNet: Imperceptible and blackbox-undetectable backdoors in compiled neural networks. (arXiv:2210.00108v2 [cs.LG] UPDATED)

Early backdoor attacks against machine learning set off an arms race in
attack and defence development. Defences have since appeared demonstrating some
ability to detect backdoors in models or even remove them. These defences work
by inspecting the training data, the model, or the integrity of the training
procedure. In this work, we show that backdoors can be added during
compilation, circumventing any safeguards in the data preparation and model
training stages. As an illustration, the attacker can insert weight-based …

backdoors networks neural networks undetectable