The Most Comprehensive Model Scanner
ML models are shared over the internet, between teams, and are used to make critical decisions. Despite this, models are not scanned with the rigor of a PDF file in your inbox. ModelScan is an open source project that scans models to determine if they contain unsafe code. It is the first model scanning tool to support multiple model formats, including H5, Pickle, and SavedModel formats. This protects you when using any format tool.
Protection Against Model Serialization Attacks
Models often need to move from one machine to another before deployment. The process of saving a model to it's destination is called serialization. A Model Serialization Attack is where malicious code is injected to the contents of a model during that process. ModelScan helps protect you from various types of Model Serialization Attacks.
Cloud credentials for writing and reading data to other systems in your environment are at risk of being compromised.
The request during the serialization process that is sent to the model is exposed to theft of the data used to build the model.
The data sent after the model performs it's task can be poisoned, opening you up to security risks.
If an attack takes place, the results of the model itself can be poisoned, which has a downstream impact on the AI application it is powering.