METHODS AND SYSTEMS FOR NEURAL PROCESSOR TRAINING BY ENCOURAGEMENT OF CORRECT OUTPUT
Patent №
US 8,892,485
Granted
2014-11-18
Filed 2010
Owner
QUALCOMM INCORPORATED
Lab
—
AI components
3
ml · vision · hardware
Assignment
Recorded
Dataset
AIPD
2023_r1 edition
Application
12832399
Certain embodiments of the present disclosure support implementation of a neural processor with synaptic weights, wherein training of the synapse weights is based on encouraging a specific output neuron to generate a spike. The implemented neural processor can be applied for classification of images and other patterns.
AI classification
Vision1.00
Machine learning1.00
AI hardware1.00
Knowledge representation0.02
Natural language0.01
Planning0.01
Evolutionary computation0.01
Speech0.00
Ownership
QUALCOMM INCORPORATED
assignment · 248370393
Assignors
APARIN, VLADIMIR, LEVIN, JEFFREY A.
On an employer assignment, the assignors are typically the inventors.
From the same owner
NEURAL NETWORK OR LAYER CONFIGURATION INDICATOR FOR A CHANNEL STATE INFORMATION SCHEMEUS 12,694,289 · 2026DEPTH ESTIMATION BASED ON FEATURE RECONSTRUCTION WITH ADAPTIVE MASKING AND MOTION PREDICTIONUS 12,694,550 · 2026STRUCTURE-FROM-MOTION (SFM) SEQUENCING USING NON-VISUAL INFORMATIONUS 12,694,610 · 2026MESSAGE PASSING NETWORK BASED OBJECT SIGNATURE FOR OBJECT TRACKINGUS 12,694,647 · 2026SYSTEMS AND METHODS FOR SEGMENTATION MAP ERROR CORRECTIONUS 12,694,667 · 2026INDICATING CAUSES FOR LIFE CYCLE MANAGEMENT OPERATIONSUS 12,696,107 · 2026SPECULATIVE DECODING IN AUTOREGRESSIVE GENERATIVE ARTIFICIAL INTELLIGENCE MODELSUS 12,688,235 · 2026DISTRIBUTED GENERATION OF VIRTUAL CONTENTUS 12,688,650 · 2026