Assignee

PERCEIVE CORP

US91 patents

Top patents by PatentIndex Score

US11250326B1Feb 15, 2022

Splitting neural network filters for implementation by neural network inference circuit

PERCEIVE CORP50 citations98

US11170289B1Nov 9, 2021

Compiler for implementing memory shutdown for neural network implementation configuration

PERCEIVE CORP25 citations95

US11049013B1Jun 29, 2021

Encoding of weight values stored on neural network inference circuit

PERCEIVE CORP39 citations95

US11610154B1Mar 21, 2023

Preventing overfitting of hyperparameters during training of network

PERCEIVE CORP22 citations94

US11468145B1Oct 11, 2022

Storage of input values within core of neural network inference circuit

PERCEIVE CORP21 citations94

US11347297B1May 31, 2022

Neural network inference circuit employing dynamic memory sleep

PERCEIVE CORP15 citations94

US11210586B1Dec 28, 2021

Weight value decoder of neural network inference circuit

PERCEIVE CORP25 citations94

US11151695B1Oct 19, 2021

Video denoising using neural networks with spatial and temporal features

PERCEIVE CORP13 citations94

US10863127B1Dec 8, 2020

Compressive sensing based image capture using multi-lens array

PERCEIVE CORP12 citations94

US12165069B1Dec 10, 2024

Compiler for optimizing number of cores used to implement neural network

PERCEIVE CORP4 citations86

US12136039B1Nov 5, 2024

Optimizing global sparsity for neural network

PERCEIVE CORP12 citations86

US12061988B1Aug 13, 2024

Decomposition of ternary weight tensors

PERCEIVE CORP6 citations86

US11625585B1Apr 11, 2023

Compiler for optimizing filter sparsity for neural network implementation configuration

PERCEIVE CORP7 citations86

US11604973B1Mar 14, 2023

Replication of neural network layers

PERCEIVE CORP11 citations86

US11586910B1Feb 21, 2023

Write cache for neural network inference circuit

PERCEIVE CORP8 citations86

US11568227B1Jan 31, 2023

Neural network inference circuit read controller with multiple operational modes

PERCEIVE CORP14 citations86

US11537870B1Dec 27, 2022

Training sparse networks with discrete weight values

PERCEIVE CORP8 citations86

US11531879B1Dec 20, 2022

Iterative transfer of machine-trained network inputs from validation set to training set

PERCEIVE CORP16 citations86

US11531868B1Dec 20, 2022

Input value cache for temporarily storing input values

PERCEIVE CORP6 citations86

US11222257B1Jan 11, 2022

Non-dot product computations on neural network inference circuit

PERCEIVE CORP8 citations86

US10586151B1Mar 10, 2020

Mitigating overfitting in training machine trained networks

PERCEIVE CORP15 citations86

US11847567B1Dec 19, 2023

Loss-aware replication of neural network layers

PERCEIVE CORP11 citations85

US11475310B1Oct 18, 2022

Training network to minimize worst-case error

PERCEIVE CORP8 citations85

US11531727B1Dec 20, 2022

Computation of neural network node with large input values

PERCEIVE CORP3 citations84

US11501138B1Nov 15, 2022

Control circuits for neural network inference circuit

PERCEIVE CORP3 citations84

US11403530B1Aug 2, 2022

Using quinary weights with neural network inference circuit designed for ternary weights

PERCEIVE CORP4 citations84

US11295200B1Apr 5, 2022

Time-multiplexed dot products for neural network inference circuit

PERCEIVE CORP4 citations84

US11205115B1Dec 21, 2021

Neural network inference circuit

PERCEIVE CORP5 citations84

US11163986B2Nov 2, 2021

Using batches of training items for training a network

PERCEIVE CORP4 citations84

US11113603B1Sep 7, 2021

Training network with discrete weight values

PERCEIVE CORP5 citations84

US11017295B1May 25, 2021

Device storing ternary weight parameters for machine-trained network

PERCEIVE CORP4 citations84

US10977338B1Apr 13, 2021

Reduced-area circuit for dot product computation

PERCEIVE CORP4 citations84

US10867247B1Dec 15, 2020

Machine learning through multiple layers of novel machine trained processing nodes

PERCEIVE CORP10 citations84

US10671888B1Jun 2, 2020

Using batches of training items for training a network

PERCEIVE CORP6 citations84

US11043006B1Jun 22, 2021

Use of machine-trained network for misalignment identification

PERCEIVE CORP4 citations83

US10742959B1Aug 11, 2020

Use of machine-trained network for misalignment-insensitive depth perception

PERCEIVE CORP7 citations83

US12112254B1Oct 8, 2024

Optimizing loss function during training of network

PERCEIVE CORP5 citations75

US12045725B1Jul 23, 2024

Batch normalization for replicated layers of neural network

PERCEIVE CORP5 citations75

US11900238B1Feb 13, 2024

Removing nodes from machine-trained network based on introduction of probabilistic noise during training

PERCEIVE CORP6 citations75

US11886979B1Jan 30, 2024

Shifting input values within input buffer of neural network inference circuit

PERCEIVE CORP3 citations75

US11783167B1Oct 10, 2023

Data transfer for non-dot product computations on neural network inference circuit

PERCEIVE CORP5 citations75

US12001948B1Jun 4, 2024

Machine trained network using novel coding techniques

PERCEIVE CORP3 citations74

US12093696B1Sep 17, 2024

Bus for transporting output values of a neural network layer to cores specified by configuration data

PERCEIVE CORP2 citations73

US12093816B1Sep 17, 2024

Initialization of values for training a neural network with quantized weights

PERCEIVE CORP5 citations73

US12061981B1Aug 13, 2024

Decomposition of weight tensors in network with value quantization

PERCEIVE CORP2 citations73

US11995555B1May 28, 2024

Training a neural network with quantized weights

PERCEIVE CORP3 citations73

Showing the top 50 of 91 patents by PatentIndex Score.

PERCEIVE CORP

Top patents by PatentIndex Score

Splitting neural network filters for implementation by neural network inference circuit

Computation of neural network node by neural network inference circuit

Reduced dot product computation circuit

Probabilistic loss function for training network with triplets

Machine-trained network for misalignment-insensitive depth perception

Compiler for implementing memory shutdown for neural network implementation configuration

Encoding of weight values stored on neural network inference circuit

Preventing overfitting of hyperparameters during training of network

Storage of input values within core of neural network inference circuit

Neural network inference circuit employing dynamic memory sleep

Weight value decoder of neural network inference circuit

Video denoising using neural networks with spatial and temporal features

Compressive sensing based image capture using multi-lens array

Compiler for optimizing number of cores used to implement neural network

Optimizing global sparsity for neural network

Decomposition of ternary weight tensors

Compiler for optimizing filter sparsity for neural network implementation configuration

Replication of neural network layers

Write cache for neural network inference circuit

Neural network inference circuit read controller with multiple operational modes

Training sparse networks with discrete weight values

Iterative transfer of machine-trained network inputs from validation set to training set

Input value cache for temporarily storing input values

Non-dot product computations on neural network inference circuit

Mitigating overfitting in training machine trained networks

Loss-aware replication of neural network layers

Training network to minimize worst-case error

Computation of neural network node with large input values

Control circuits for neural network inference circuit

Using quinary weights with neural network inference circuit designed for ternary weights

Time-multiplexed dot products for neural network inference circuit

Neural network inference circuit

Using batches of training items for training a network

Training network with discrete weight values

Device storing ternary weight parameters for machine-trained network

Reduced-area circuit for dot product computation

Machine learning through multiple layers of novel machine trained processing nodes

Using batches of training items for training a network

Use of machine-trained network for misalignment identification

Use of machine-trained network for misalignment-insensitive depth perception

Optimizing loss function during training of network

Batch normalization for replicated layers of neural network

Removing nodes from machine-trained network based on introduction of probabilistic noise during training

Shifting input values within input buffer of neural network inference circuit

Data transfer for non-dot product computations on neural network inference circuit

Machine trained network using novel coding techniques

Bus for transporting output values of a neural network layer to cores specified by configuration data

Initialization of values for training a neural network with quantized weights

Decomposition of weight tensors in network with value quantization

Training a neural network with quantized weights