Patents by Inventor Zhifeng Chen
Zhifeng Chen has filed for patents to protect the following inventions. This listing includes patent applications that are pending as well as patents that have already been granted by the United States Patent and Trademark Office (USPTO).
-
Patent number: 11450120Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for processing point cloud data representing a sensor measurement of a scene captured by one or more sensors to generate an object detection output that identifies locations of one or more objects in the scene. When deployed within an on-board system of a vehicle, the object detection output that is generated can be used to make autonomous driving decisions for the vehicle with enhanced accuracy.Type: GrantFiled: July 8, 2020Date of Patent: September 20, 2022Assignee: Waymo LLCInventors: Jonathon Shlens, Patrick An Phu Nguyen, Benjamin James Caine, Jiquan Ngiam, Wei Han, Brandon Chauloon Yang, Yuning Chai, Pei Sun, Yin Zhou, Xi Yi, Ouais Alsharif, Zhifeng Chen, Vijay Vasudevan
-
Publication number: 20220180193Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for training a neural network to perform 3D object detection. One of the methods includes training a student neural network to perform 3D object detection using pseudo-labels generated by a teacher neural network.Type: ApplicationFiled: December 9, 2021Publication date: June 9, 2022Inventors: Benjamin James Caine, Rebecca Dawn Roelofs, Jonathon Shlens, Zhifeng Chen, Jiquan Ngiam, Vijay Vasudevan
-
Publication number: 20220178254Abstract: A cutting device includes a base, an operation bench, and a cutting mechanism. The base includes a fluid cavity for containing a fluid. The operation bench is disposed on the base. The cutting mechanism includes a driving member and a cutting member. The driving member drives the cutting member to rotate. The cutting member at least partially protrudes from and passes through the operation bench. The cutting device includes a fluid control mechanism including a radial flow guiding member. The radial flow guiding member is disposed in the fluid cavity and includes at least a flow guiding surface disposed around a periphery of the cutting member. The fluid is capable of flowing along a surface of the flow guiding surface and the flow guiding surface is inclined or curved with respect to a bottom wall of the fluid cavity.Type: ApplicationFiled: December 6, 2021Publication date: June 9, 2022Inventors: Zhifeng Chen, Dianbo Zhu, Huaping Pan, Changning Zhang
-
Patent number: 11351640Abstract: Provided is a cutting tool that includes a cutting head, a frame, a working table, a guiding rail assembly, and a fence. The cutting head is configured to perform a cutting function. The frame is configured to support the cutting head. The working table is configured to place an object to be cut. The guiding rail assembly is configured to enable the working table to slide along a first straight line relative to the frame. The fence is mounted to the working table, includes at least one fence surface extending in a direction perpendicular to the first straight line. The cutting tool enables a workpiece to be cut to be stably placed on the working table during a cutting operation.Type: GrantFiled: December 13, 2018Date of Patent: June 7, 2022Assignee: Nanjing Chervon Industry Co., Ltd.Inventors: Zhifeng Chen, Zhigang Pan, Yang Cao, Yu Xu
-
Patent number: 11356638Abstract: A device may control a video communication via transcoding and/or traffic shaping. The device may include a multipoint control unit (MCU) and/or a server. The device may receive one or more video streams from one or more devices. The device may analyze a received video stream to determine a viewing parameter. The viewing parameter may include a user viewing parameter, a device viewing parameter, and/or a content viewing parameter. The device may modify a video stream based on the viewing parameter. Modifying the video stream may include re-encoding the video stream, adjusting an orientation, removing a video detail, and/or adjusting a bit rate. The device may send the modified video stream to another device. The device may determine a hit rate for the video stream based on the viewing parameter. The device may indicate the bit rate by sending a feedback message and/or by signaling a bandwidth limit.Type: GrantFiled: April 19, 2018Date of Patent: June 7, 2022Assignee: VID SCALE, Inc.Inventors: Yuriy Reznik, Gregory S. Sternberg, Ariela Zeira, Liangping Ma, Yong He, Anantharaman Balasubramanian, Yan Ye, Louis Kerofsky, Ralph Neff, Zhifeng Chen
-
Publication number: 20220152760Abstract: A power device includes a power tool, a mounting plate, a mounting table, a positioning assembly, and a locking assembly. The positioning assembly includes a positioning piece. A positioning space is formed between the positioning piece and the mounting table. An end of the mounting plate is enabled to be inserted into the positioning space so that the mounting plate is pre-positioned. The locking assembly includes a movable piece and a locking piece. The movable piece includes a support position where the mounting plate is supported and an unlocked position where support for the mounting plate is released. The locking piece fixes at least the mounting plate to the mounting table.Type: ApplicationFiled: November 15, 2021Publication date: May 19, 2022Inventors: Zhifeng Chen, Dianbo Zhu, Bo Zhang, Xia Wang
-
Publication number: 20220130374Abstract: Methods, systems, and apparatus, including computer programs encoded on a computer-readable media, for speech recognition using multi-dialect and multilingual models. In some implementations, audio data indicating audio characteristics of an utterance is received. Input features determined based on the audio data are provided to a speech recognition model that has been trained to output score indicating the likelihood of linguistic units for each of multiple different language or dialects. The speech recognition model can be one that has been trained using cluster adaptive training. Output that the speech recognition model generated in response to receiving the input features determined based on the audio data is received. A transcription of the utterance generated based on the output of the speech recognition model is provided.Type: ApplicationFiled: January 10, 2022Publication date: April 28, 2022Applicant: Google LLCInventors: Zhifeng Chen, Bo Li, Eugene Weinstein, Yonghui Wu, Pedro J. Moreno Mengibar, Ron J. Weiss, Khe Chai Sim, Tara N. Sainath, Patrick An Phu Nguyen
-
Publication number: 20220121945Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for training giant neural networks. One of the methods includes obtaining data specifying a partitioning of the neural network into N composite layers that form a sequence of composite layers, wherein each composite layer comprises a distinct plurality of layers from the multiple network layers of the neural network; obtaining data assigning each of the N composite layers to one or more computing devices from a set of N computing devices; partitioning a mini-batch of training examples into a plurality of micro-batches; and training the neural network, comprising: performing a forward pass through the neural network until output activations have been computed for each micro-batch for a final composite layer in the sequence, and performing a backward pass through the neural network until output gradients have been computed for each micro-batch for the first composite layer in the sequence.Type: ApplicationFiled: January 3, 2022Publication date: April 21, 2022Inventors: Zhifeng Chen, Yanping Huang, Youlong Cheng, HyoukJoong Lee, Dehao Chen, Jiquan Ngiam
-
Publication number: 20220083746Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for machine translation using neural networks. In some implementations, a text in one language is translated into a second language using a neural network model. The model can include an encoder neural network comprising a plurality of bidirectional recurrent neural network layers. The encoding vectors are processed using a multi-headed attention module configured to generate multiple attention context vectors for each encoding vector. A decoder neural network generates a sequence of decoder output vectors using the attention context vectors. The decoder output vectors can represent distributions over various language elements of the second language, allowing a translation of the text into the second language to be determined based on the sequence of decoder output vectors.Type: ApplicationFiled: August 27, 2021Publication date: March 17, 2022Inventors: Zhifeng Chen, Macduff Richard Hughes, Yonghui Wu, Michael Schuster, Xu Chen, Llion Owen Jones, Niki J. Parmar, George Foster, Orhan Firat, Ankur Bapna, Wolfgang Macherey, Melvin Jose Johnson Premkumar
-
Publication number: 20220086047Abstract: A SDN orchestration method includes: obtaining a first request for creating a first logical switch; creating a control plane instance of the first logical switch, and sending first configuration information to instruct the first forwarding device to configure the data plane instance of the first logical switch; obtaining a second request for connecting the first logical switch to a first logical router; sending second configuration information to instruct the first forwarding device to configure a first port of the data plane instance of the first logical switch to be communicatively connected to a second port of a data plane instance of the first logical router on the second forwarding device configured with the data plane instance of the first logical router; and sending third configuration information to instruct the second forwarding device to configure the second port to be communicatively connected to the first port.Type: ApplicationFiled: September 27, 2021Publication date: March 17, 2022Inventors: Zhifeng CHEN, Xuefeng WU, Weisheng WANG, Chenghao LI
-
Publication number: 20220034902Abstract: A method is described for identifying high affinity monoclonal antibody heavy and light chain pairs from high throughput screens of antibody producing cell libraries such as B-cell and hybridoma libraries. Specifically, the method relates to application of reversed immunocapture and high resolution tandem mass spectrometry for the identification of heavy and light chain pairs of binding antibodies obtained from high throughput screens of antibody producing cell libraries.Type: ApplicationFiled: December 6, 2019Publication date: February 3, 2022Applicant: Merck Sharp & Dohme Corp.Inventors: Eberhard Durr, Yaping Liu, Zhifeng Chen, Arthur Fridman
-
Patent number: 11238845Abstract: Methods, systems, and apparatus, including computer programs encoded on a computer-readable media, for speech recognition using multi-dialect and multilingual models. In some implementations, audio data indicating audio characteristics of an utterance is received. Input features determined based on the audio data are provided to a speech recognition model that has been trained to output score indicating the likelihood of linguistic units for each of multiple different language or dialects. The speech recognition model can be one that has been trained using cluster adaptive training. Output that the speech recognition model generated in response to receiving the input features determined based on the audio data is received. A transcription of the utterance generated based on the output of the speech recognition model is provided.Type: GrantFiled: November 14, 2019Date of Patent: February 1, 2022Assignee: Google LLCInventors: Zhifeng Chen, Bo Li, Eugene Weinstein, Yonghui Wu, Pedro J. Moreno Mengibar, Ron J. Weiss, Khe Chai Sim, Tara N. Sainath, Patrick An Phu Nguyen
-
Patent number: 11232356Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for training giant neural networks. One of the methods includes obtaining data specifying a partitioning of the neural network into N composite layers that form a sequence of composite layers, wherein each composite layer comprises a distinct plurality of layers from the multiple network layers of the neural network; obtaining data assigning each of the N composite layers to one or more computing devices from a set of N computing devices; partitioning a mini-batch of training examples into a plurality of micro-batches; and training the neural network, comprising: performing a forward pass through the neural network until output activations have been computed for each micro-batch for a final composite layer in the sequence, and performing a backward pass through the neural network until output gradients have been computed for each micro-batch for the first composite layer in the sequence.Type: GrantFiled: August 10, 2020Date of Patent: January 25, 2022Assignee: Google LLCInventors: Zhifeng Chen, Yanping Huang, Youlong Cheng, HyoukJoong Lee, Dehao Chen, Jiquan Ngiam
-
Publication number: 20220005465Abstract: A method for performing speech recognition using sequence-to-sequence models includes receiving audio data for an utterance and providing features indicative of acoustic characteristics of the utterance as input to an encoder. The method also includes processing an output of the encoder using an attender to generate a context vector, generating speech recognition scores using the context vector and a decoder trained using a training process, and generating a transcription for the utterance using word elements selected based on the speech recognition scores. The transcription is provided as an output of the ASR system.Type: ApplicationFiled: September 20, 2021Publication date: January 6, 2022Applicant: Google LLCInventors: Rohit Prakash Prabhavalkar, Zhifeng Chen, Bo Li, Chung-cheng Chiu, Kanury Kanishka Rao, Yonghui Wu, Ron J. Weiss, Navdeep Jaitly, Michiel A.u. Bacchiani, Tara N. Sainath, Jan Kazimierz Chorowski, Anjuli Patricia Kannan, Ekaterina Gonina, Patrick An Phu Nguyen
-
Patent number: 11210272Abstract: An example distributed database includes a first instance and a second instance. The first instance is configured to: responsive to performing, within a scope of a database update transaction, a first database update operation, invalidate a cache entry residing in the first database cache maintained by the first instance, wherein the first database update operation is reflected by a transaction log maintained by the first instance; perform, within the scope of the database update transaction, a second database update operation to insert an identifier of the cache entry into a predetermined table of the distributed database, wherein the second database update operation is reflected by the transaction log; and responsive to committing the database update transaction, transmit the transaction log to the second instance.Type: GrantFiled: April 25, 2019Date of Patent: December 28, 2021Assignee: Electronic Arts Inc.Inventors: Xinfeng Zhang, Mengxin Ye, Zhifeng Chen, Xiaokai Wu
-
Publication number: 20210390271Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for neural machine translation. The method comprises obtaining a first sequence of words in a source language, generating a modified sequence of words in the source language by inserting a word boundary symbol only at the beginning of each word in the first sequence of words and not at the end of each word, dividing the modified sequence of words into wordpieces using a wordpiece model, generating, from the wordpieces, an input sequence of input tokens for a neural machine translation system; and generating an output sequence of words using the neural machine translation system based on the input sequence of input tokens.Type: ApplicationFiled: August 27, 2021Publication date: December 16, 2021Inventors: Mohammad Norouzi, Zhifeng Chen, Yonghui Wu, Michael Schuster, Quoc V. Le
-
Publication number: 20210366463Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for generating speech from text. One of the systems includes one or more computers and one or more storage devices storing instructions that when executed by one or more computers cause the one or more computers to implement: a sequence-to-sequence recurrent neural network configured to: receive a sequence of characters in a particular natural language, and process the sequence of characters to generate a spectrogram of a verbal utterance of the sequence of characters in the particular natural language; and a subsystem configured to: receive the sequence of characters in the particular natural language, and provide the sequence of characters as input to the sequence-to-sequence recurrent neural network to obtain as output the spectrogram of the verbal utterance of the sequence of characters in the particular natural language.Type: ApplicationFiled: August 2, 2021Publication date: November 25, 2021Inventors: Samuel Bengio, Yuxuan Wang, Zongheng Yang, Zhifeng Chen, Yonghui Wu, Ioannis Agiomyrgiannakis, Ron J. Weiss, Navdeep Jaitly, Ryan M. Rifkin, Robert Andrew James Clark, Quoc V. Le, Russell J. Ryan, Ying Xiao
-
Publication number: 20210358491Abstract: Methods, systems, and apparatus, including computer programs encoded on computer-readable storage media, for speech recognition using attention-based sequence-to-sequence models. In some implementations, audio data indicating acoustic characteristics of an utterance is received. A sequence of feature vectors indicative of the acoustic characteristics of the utterance is generated. The sequence of feature vectors is processed using a speech recognition model that has been trained using a loss function that uses N-best lists of decoded hypotheses, the speech recognition model including an encoder, an attention module, and a decoder. The encoder and decoder each include one or more recurrent neural network layers. A sequence of output vectors representing distributions over a predetermined set of linguistic units is obtained. A transcription for the utterance is obtained based on the sequence of output vectors. Data indicating the transcription of the utterance is provided.Type: ApplicationFiled: July 27, 2021Publication date: November 18, 2021Applicant: Google LLCInventors: Rohit Prakash Prabhavalkar, Tara N. Sainath, Yonghui Wu, Patrick An Phu Nguyen, Zhifeng Chen, Chung-Cheng Chiu, Anjuli Patricia Kannan
-
Patent number: 11146451Abstract: A SDN orchestration method includes: obtaining a first request for creating a first logical switch; creating a control plane instance of the first logical switch, and sending first configuration information to instruct the first forwarding device to configure the data plane instance of the first logical switch; obtaining a second request for connecting the first logical switch to a first logical router; sending second configuration information to instruct the first forwarding device to configure a first port of the data plane instance of the first logical switch to be communicatively connected to a second port of a data plane instance of the first logical router on the second forwarding device configured with the data plane instance of the first logical router; and sending third configuration information to instruct the second forwarding device to configure the second port to be communicatively connected to the first port.Type: GrantFiled: November 27, 2019Date of Patent: October 12, 2021Assignee: HUAWEI TECHNOLOGIES CO., LTD.Inventors: Zhifeng Chen, Xuefeng Wu, Weisheng Wang, Chenghao Li
-
Patent number: 11145293Abstract: Methods, systems, and apparatus, including computer-readable media, for performing speech recognition using sequence-to-sequence models. An automated speech recognition (ASR) system receives audio data for an utterance and provides features indicative of acoustic characteristics of the utterance as input to an encoder. The system processes an output of the encoder using an attender to generate a context vector and generates speech recognition scores using the context vector and a decoder trained using a training process that selects at least one input to the decoder with a predetermined probability. An input to the decoder during training is selected between input data based on a known value for an element in a training example, and input data based on an output of the decoder for the element in the training example. A transcription is generated for the utterance using word elements selected based on the speech recognition scores. The transcription is provided as an output of the ASR system.Type: GrantFiled: July 19, 2019Date of Patent: October 12, 2021Assignee: Google LLCInventors: Rohit Prakash Prabhavalkar, Zhifeng Chen, Bo Li, Chung-Cheng Chiu, Kanury Kanishka Rao, Yonghui Wu, Ron J. Weiss, Navdeep Jaitly, Michiel A. U. Bacchiani, Tara N. Sainath, Jan Kazimierz Chorowski, Anjuli Patricia Kannan, Ekaterina Gonina, Patrick An Phu Nguyen