Tag: hybrid machine learning

New Attack Technique ‘Sleepy Pickle’ Targets Machine Learning Models
News

New Attack Technique ‘Sleepy Pickle’ Targets Machine Learning Models

The discovery of a new "hybrid machine learning (ML) model exploitation technique" called Sleepy Pickle has brought the security issues associated with the Pickle format to light once again. According to Trail of Bits, the attack technique weaponizes the widely used format for distributing and packaging machine learning (ML) models in order to tamper with the model itself, seriously endangering the downstream clients of an organization's supply chain. Security researcher Boyan Milanov described Sleepy Pickle as "a stealthy and novel attack technique that targets the ML model itself rather than the underlying system." Pickle is a popular serialization format used by machine learning tools such as PyTorch read more New Attack Technique 'Sleepy Pickle' Targets Machine Learning Model...