HiddenLayer has formed a strategic partnership with Intel to provide machine learning (ML) security solutions for regulatory and compliance-focused businesses.
The two companies will work to provide customers with a hardware- and software-based end-to-end ML model security platform by identifying existing gaps in the concept of ML scanning in a secure environment.
This includes the integration of Confidential Computing scale, performance, and privacy capabilities developed using Intel’s Software Guard Extensions (SGX) and HiddenLayer’s adversarial ML threat protection capabilities.
Confidential Computing gives customers access to an encrypted ML Model Scanner to protect their respective ML models.
Deployed to power confidential computing technologies, Intel SGX supports the adoption of AI-based technologies in various industries.
Rick Echevarria, vice president and general manager of security for the Intel Sales and Marketing Group, said: A.I. ”
With the introduction of the new ML Model Scanner, customers will be able to perform comprehensive detections to recognize corruption issues.
The scanner also enables organizations with strict regulatory and compliance requirements to perform malware analysis, vulnerability assessment, and model integrity checking in a trusted execution environment.
This allows customers to protect their intellectual property, including sensitive data and AI.
HiddenLayer says its platform can support multiple ML model file types including Pickle, ONNX, Dill, Numpy, Joblib, HDF5 and Zip.
HiddenLayer CEO and Co-Founder Christopher Sestito said:
