Sentinel gate for modulating auxiliary information in a long short-term memory (lstm) neural network

LU Jiasen, C **ong, R Socher - US Patent 10,565,306, 2020 - Google Patents
The technology disclosed presents a novel spatial attention model that uses current hidden
state information of a decoder long short-term memory (LSTM) to guide attention and to …

Spatial attention model for image captioning

LU Jiasen, C **ong, R Socher - US Patent 10,558,750, 2020 - Google Patents
The technology disclosed presents a novel spatial attention model that uses current hidden
state information of a decoder long short-term memory (LSTM) to guide attention and to …

Dense video captioning

Y Zhou, L Zhou, C **ong, R Socher - US Patent 10,542,270, 2020 - Google Patents
Systems and methods for dense captioning of a video include a multi-layer encoder stack
configured to receive information extracted from a plurality of video frames, a proposal …

Adaptive attention model for image captioning

LU Jiasen, C **ong, R Socher - US Patent 10,565,305, 2020 - Google Patents
The technology disclosed presents a novel spatial attention model that uses current hidden
state information of a decoder long short-term memory (LSTM) to guide attention and to …

Natural language processing using a neural network

B McCann, C **ong, R Socher - US Patent 10,699,060, 2020 - Google Patents
A system includes a neural network for performing a first natural language processing task.
The neural network includes a first rectifier linear unit capable of executing an activation …

Natural language processing using context specific word vectors

B McCann, C **ong, R Socher - US Patent 10,817,650, 2020 - Google Patents
A system is provided for natural language processing. In some embodiments, the system
includes an encoder for generating context-specific word vectors for at least one input …

Block-diagonal hessian-free optimization for recurrent and convolutional neural networks

H Zhang, C **ong - US Patent 11,386,327, 2022 - Google Patents
Embodiments for training a neural network are provided. A neural network is divided into a
first block and a second block, and the parameters in the first block and second block are …

Question answering from minimal context over documents

MIN Sewon, V Zhong, C **ong, R Socher - US Patent 11,227,218, 2022 - Google Patents
A natural language processing system that includes a sen tence selector and a question
answering module. The sen tence selector receives a question and sentences that are …

Systems and methods for learning for domain adaptation

E Hosseini-Asl, C **ong, Y Zhou, R Socher - US Patent 11,106,182, 2021 - Google Patents
A method for training parameters of a first domain adaptation model includes evaluating a
cycle consistency objective using a first task specific model associated with a first domain …

Unsupervised non-parallel speech domain adaptation using a multi-discriminator adversarial network

E Hosseini-Asl, C **ong, Y Zhou, R Socher - US Patent 10,783,875, 2020 - Google Patents
A system for domain adaptation includes a domain adaptation model configured to adapt a
representation of a signal in a first domain to a second domain to generate an adapted …