Patent Leaderboard
USPTO Patent Rankings Data through Dec 31, 2025
Home › Constraining function approximation hardware integrated with fixed-point to floating-point conversion

Constraining function approximation hardware integrated with fixed-point to floating-point conversion

US Patent 11669747 · Granted Jun 6, 2023

Estimated economic value: $9,942,000 ⓘ

Assignee

Inventors

View full patent text on Google Patents →

More Patents by These Inventors

Simulated low bit-width quantization using bit shifted neural network parameters2025
Depth-first deep convolutional neural network inference2025
Single instruction multiple data (SIMD) sparse decompression with variable densi2025
Single instruction multiple data (SIMD) sparse decompression with variable densi2025
Single instruction multiple data (SIMD) sparse decompression with variable densi2025
Variable quantization for neural networks2025
Processor instruction for dynamic floating point exponent extraction2025
Processor instruction for dynamic floating point exponent extraction2025
Configuring arbitrary jumps in a linear feedback shift register in wireless comm2025
Compression of sparse deep convolutional network weights2025
Compression of sparse deep convolutional network weights2025
Exploiting activation sparsity in deep neural networks2024