https://store-images.s-microsoft.com/image/apps.14289.28270a09-dece-461a-aa97-bea2464688fb.ab9c1de9-e275-40cc-82c2-9aa43be379ca.9d1c76db-f22b-4b30-b5d5-f877fe6b187d

HiddenLayer Model Scanner

HiddenLayer, Inc

HiddenLayer Model Scanner

HiddenLayer, Inc

Detect malicious code and ensure your models are secure

HiddenLayer Model Scanner prevents security issues by detecting malicious code in your AI models and ensures your AI models are free from adversarial code.


HiddenLayer’s Model Scanner ensures models are free from adversarial code before entering corporate environments. The HiddenLayer Model Scanner allows data science teams to access AI models securely, allowing faster time to model deployment. The HiddenLayer Model Scanner is available via SaaS or on-prem. The Model Scanner integrates seamlessly with modern ML development lifecycles and SecOps workflows. In this era of AI innovation, the Model Scanner is a critical component for rapid collaboration and innovation. Trust, flexibility, and comprehensiveness are non-negotiable when it comes to ensuring your business stays ahead in innovation.

Key Benefits:


Powerful: Scan a broader range of model file types across more MLOps platform scenarios than any other scanner.
Flexible: Deploy on-premise and/or SaaS for seamless support in demo, training, and production environments.
Recognized: Industry-backed and award-winning, with accolades from RSA Sandbox, M12, IBM, DoD, and MITRE ATLAS alignment for integration into SecOps workflows.

Key Capabilities:


• Malware Analysis — Scans AI Models for embedded malicious code
that could serve as an infection vector & launchpad for malware
• Model Integrity — Analysis of AI Model’s layers, components & tensors to detect tampering or corruption.
• Uses a combination of static detection and analysis to identify malware, vulnerabilities, model integrity & corruption issues
• Catalog a Known-Good State of your AI Models as a baseline for identifying future tampering
• Supports ChatGPT and LLMs
• Data Leakage — Ensure LLM outputs do not expose backend systems risking privilege escalation or remove code execution.
• Supports a variety of AI Model file types: Pickle, Dill, Joblib, Numpy, Zip, and ONNX
https://store-images.s-microsoft.com/image/apps.58131.28270a09-dece-461a-aa97-bea2464688fb.b486d247-c494-48ad-b314-60fbff7f2b75.4528bd43-41b8-46e9-9b91-bfdbcf9ed3f6
https://store-images.s-microsoft.com/image/apps.58131.28270a09-dece-461a-aa97-bea2464688fb.b486d247-c494-48ad-b314-60fbff7f2b75.4528bd43-41b8-46e9-9b91-bfdbcf9ed3f6
https://store-images.s-microsoft.com/image/apps.53536.28270a09-dece-461a-aa97-bea2464688fb.b486d247-c494-48ad-b314-60fbff7f2b75.f3708bc3-66e7-485d-a6ad-ee9b3b3ab801