CryptoKANs+: KAN-Inspired Self-Learning Polynomial Networks for Efficient Privacy-Preserving Machine Learning

Computer Science

• Activation Function 50%
• Encrypted Data 50%
• Homomorphic Encryption 50%
• Linear Transformation 50%
• Multilayer Perceptron 50%
• Network Inference 50%
• Neural Network 50%
• Performance Degradation 50%
• Polynomial Approximation 50%
• Privacy Concern 50%
• Privacy Preserving Machine Learning 100%

Keyphrases

• Cryptonets 25%
• Encrypted Inference 75%
• Kolmogorov-Arnold Networks 100%
• Learned Activations 25%
• Multiplicative Depth 25%
• Network Inference 25%
• Polynomial Activation Function 75%
• Polynomial Networks 100%
• Pre-activation 25%
• Quadratic Transformation 25%
• Transformation Layer 25%