Inventor · Stanford, CA, US

Quoc V. Le

112Patents
8h-index
167Co-inventors
79Inventor score

Filing activity: Mar 14, 2013 → Dec 3, 2024

Most-cited inventions

PatentTitleAreaCited byStatus
US9799327B1 Speech recognition with attention-based recurrent neural networks Physics 46 Active
US9256807B1 Generating labeled images Physics 23 Active
US10528866B1 Training a document classification neural network Physics 17 Active
US9984062B1 Generating author vectors Physics 15 Active
US10181098B2 Generating representations of input sequences using neural networks Physics 14 Active
US10997503B2 Computationally efficient neural network architecture search Physics 11 Active
US9990918B1 Speech recognition with attention-based recurrent neural networks Physics 9 Active
US9805028B1 Translating terms using numeric representations Physics 8 Active
US10963779B2 Neural network programmer Physics 8 Active
US10043512B2 Generating target sequences from input sequences using partial conditioning Physics 8 Active
US10713593B2 Implicit bridging of machine learning tasks Physics 8 Active
US10133739B2 Neural machine translation systems with rare word processing Physics 6 Active
US10803380B2 Generating vector representations of documents Physics 5 Active
US11164066B1 Generating parameter values for recurrent neural networks Physics 5 Active
US10366327B2 Generating vector representations of documents Physics 4 Active
US11030523B2 Neural architecture search Physics 4 Active
US11200492B1 Training a document classification neural network Physics 4 Active
US11182566B2 Processing text sequences using neural networks Physics 4 Active
US10540962B1 Speech recognition with attention-based recurrent neural networks Physics 4 Active
US9652695B2 Label consistency for image analysis Physics 3 Active
US10324993B2 Predicting a search engine ranking signal value Physics 3 Active
US11144831B2 Regularized neural network architecture search Physics 3 Active
US10936828B2 Neural machine translation systems with rare word processing Physics 3 Active
US10573293B2 End-to-end text-to-speech conversion Physics 3 Active
US11205099B2 Training neural networks using data augmentation policies Physics 2 Active

Source: USPTO / EPO open patent data. Inventor disambiguation is heuristic; counts are objective bibliographic measures.