CipherExplain

CipherExplain facilitates secure, privacy-preserving model explanations for regulated industries. Key capabilities include: • SHAP explanations on encrypted inputs • Fully homomorphic encryption for data privacy • Zero plaintext exposure to vendors • Compliance with SR 11-7, EU Act, and HIPAA mandates This platform addresses the critical challenge of reconciling data privacy with transparency requirements in high-risk applications. It computes SHAP feature attributions entirely under Fully Homomorphic Encryption, ensuring that sensitive input data, model weights, and explanations remain encrypted throughout the process. Users can register their own models via a secure JSON API, maintaining complete control over their training data environments. CipherExplain provides verifiable encrypted SHAP explanations with reproducible benchmarks and signed audit artifacts. The system leverages techniques like O(d log d) coalition sampling and SIMD slot packing to optimize performance while maintaining strong cryptographic guarantees. This allows organizations to generate audit-ready explanations without compromising the confidentiality of personal or proprietary information. Ideal for banking, healthcare networks, and hiring platforms operating under strict data privacy regulations. Developers can integrate the functionality using a Python SDK, enabling secure model explainability in production environments where compliance and data protection are paramount.