logo

For the Inventor, By the Inventor

San Jose, CA, United States of America

Perceive Corporation


USPTO Granted Patents = 83

Forward Citations = 465

Loading Chart...
Loading Chart...

11 inventors (with patents filed for the assignee):

goldMedal Steven L Teig (77 out of 460 patents)

silverMedal Jung Ko (28 out of 48 patents)

bronzeMedal Kenneth Duong (28 out of 54 patents)

4.  Eric A Sather (26 out of 45 patents)

5.  Andrew C Mihal (17 out of 25 patents)

6.  Ilyas Mohammed (6 out of 285 patents)

7.  Brian Thomas (4 out of 9 patents)

8.  Alexandru F Drimbarean (3 out of 3 patents)

9.  Ryan J Cassidy (2 out of 3 patents)

10.  Justin Tantiongloc (1 out of 2 patents)

11.  Won Rhee (1 out of 4 patents)


Core Competencies:
Neural Network
Optimization
Quantization
Error Minimization
Machine Learning
Training Techniques
Dynamic Memory
Sparsity
Input Buffering
Depth Perception
Context-Sensitive Detection
Compressive Sensing

83 patents:

1. 12159214 - Buffering of neural network inputs and outputs (2024-12-03),
2. 12136039 - Optimizing global sparsity for neural network (2024-11-05),
3. 12124939 - Generation of machine-trained network instructions (2024-10-22),
4. 12118463 - Weight value decoder of neural network inference circuit (2024-10-15),
5. 12112254 - Optimizing loss function during training of network (2024-10-08),
6. 12093816 - Initialization of values for training a neural network with quantized weights (2024-09-17),
7. 12093696 - Bus for transporting output values of a neural network layer to cores specified by configuration data (2024-09-17),
8. 12061988 - Decomposition of ternary weight tensors (2024-08-13),
9. 12061981 - Decomposition of weight tensors in network with value quantization (2024-08-13),
10. 12051000 - Training network to minimize worst-case error (2024-07-30),
11. 12045725 - Batch normalization for replicated layers of neural network (2024-07-23),
12. 12008465 - Dynamic generation of data sets for training machine-trained network (2024-06-11),
13. 12001948 - Machine trained network using novel coding techniques (2024-06-04),
14. 11995555 - Training a neural network with quantized weights (2024-05-28),
15. 11995537 - Training network with batches of input instances (2024-05-28),
Perceive Corporation, based in San Jose, CA, specializes in advanced neural network optimization technologies. The company focuses on enhancing machine learning capabilities through innovative techniques such as quantization, error minimization, and dynamic memory management. By leveraging sparsity and efficient training methods, Perceive aims to improve the performance and efficiency of neural networks, making them more accessible for a variety of applications across industries. Their research contributes significantly to the evolving landscape of artificial intelligence.
Profile summary based on public USPTO records.

logo

Company Badge

USPTO Data Jan 1 1976 - Sep 29 2026

Perceive Corporation
83 Patents

#130 in San Jose, CA

#1,035 in California

#4,466 in US

#9,133 in the World

San Jose, CA (US)
San Jose, CA, United States of America

Perceive Corporation

Forward Citations = 465

11 inventors:

goldMedal Steven L Teig (77 out of 460 patents)

silverMedal Jung Ko (28 out of 48 patents)

bronzeMedal Kenneth Duong (28 out of 54 patents)

4.  Eric A Sather (26 out of 45 patents)

5.  Andrew C Mihal (17 out of 25 patents)

Please report any incorrect information to [email protected]
Loading…