Low-precision random Fourier features for memory-constrained kernel approximation

Keyphrases

• Memory-constrained 100%
• Kernel Approximation 100%
• Random Fourier Features 100%
• Low-precision 100%
• Higher Rank 60%
• Approximation Methods 40%
• Generalization Performance 40%
• Rank-one Approximation 40%
• Low-bit Quantization 40%
• Negligible Effects 20%
• Approximation Error 20%
• Benchmark Dataset 20%
• Nystrm Method 20%
• Conventional Metrics 20%

Computer Science

• Approximation (Algorithm) 100%
• Generalization Performance 40%
• Rank Approximation 40%
• Approximated Matrix 20%