Patents by Inventor Zhengyu Zhou

Zhengyu Zhou 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).

  • Publication number: 20250217864
    Abstract: A method includes receiving query data, receiving item data, initializing the query data as at least one natural language query token, and initializing the item data as at least one natural language item token. The method also includes generating a knowledge graph for the item based on the at least one natural language item token, flattening the knowledge graph for the item to generate a knowledge graph string, mapping at least one token associated with the knowledge graph string and the at least one natural language query token to an embedding vector using a matrix of parameters, and providing, to a machine learning model, the embedding vector. The method also includes receiving, from the machine learning model, a recommendation and a natural language explanation of the recommendation, and providing, to a user at a display, the recommendation and the natural language explanation of the recommendation.
    Type: Application
    Filed: December 29, 2023
    Publication date: July 3, 2025
    Inventors: ANTHONY M. COLAS, JUN ARAKI, ZHENGYU ZHOU, BINGQING WANG, ZHE FENG
  • Patent number: 12272353
    Abstract: The systems and methods described herein are related to a new speech understanding system for domain-specific voice interaction. The systems and methods described herein combine the automatic correction of an automatic speech recognition errors with a natural language understanding model in a way that optimizes the recognition and understanding of a received speech input. The systems and methods described herein may further support out-of-domain detection or domain classification by jointly learning and/or performing automatic speech recognition error correction and domain-related classification. Through the joint learning with automatic speech recognition error correction, the out-of-domain detection or domain classification may be conducted based on a plurality of possible speech recognition results with shared feature inputs and shared neural layers. The systems and methods described herein may achieve robust performance with high computational efficiency.
    Type: Grant
    Filed: September 26, 2022
    Date of Patent: April 8, 2025
    Assignee: Robert Bosch GmbH
    Inventor: Zhengyu Zhou
  • Publication number: 20250068967
    Abstract: An event analysis and prediction system is disclosed that is configured to analyze at least one event data stream from a monitored system for the purpose of predicting future events in the system, e.g., for predictive maintenance of the system. The event analysis and prediction system advantageously predicts, in real-time, such failures in the monitored system that would otherwise lead to interruptions. The event analysis and prediction system advantageously leverages a novel data generation procedure that, in essence, converts the problem of failure prediction to a classification problem, which is solved using machine learning algorithms. The event analysis and prediction system is designed to work well even when there is limited labeled data available for model training.
    Type: Application
    Filed: August 25, 2023
    Publication date: February 27, 2025
    Inventors: HyeongSik Kim, Andrew Le Clair, Zhengyu Zhou, Abhishek Saini, Pongtep Angkititrakul
  • Publication number: 20250061286
    Abstract: Systems and methods are described herein for detecting and reducing hallucinations and improving the reliability of an LLM-based QA system. Particularly, the disclosure provides a set of hallucination detection and handling approaches that address different types of hallucinations of the QA system. The hallucination detection approaches described herein include comparing a natural language answer with context information by way of sentence similarity estimation and keyword matching. The hallucination handling approaches described herein include removing hallucinated sentences from the natural language answer or regenerating the natural language answer using better context information, depending on a level of hallucination detected in the natural language answer. A hybrid framework is also provided that systematically combines the hallucination detection approaches and hallucination handling approaches into one system to achieve an optimal hallucination-reduction performance for the QA system.
    Type: Application
    Filed: August 18, 2023
    Publication date: February 20, 2025
    Inventors: Zhengyu Zhou, Arsalan Gundroo
  • Publication number: 20250014373
    Abstract: A plurality of weak label augmenters of different paradigms are integrated into a framework using robust training and negative instance filtering. A first of the augmenters extracts first weak labels from unlabeled data, a second of the augmenters extracts second weak labels from the unlabeled data. The robust training is used with an objective to downweight the probability of entities belonging to the wrong category. The first and second weak labels are filtered using an instance filter to update a high-precision training set shared by the plurality of augmenters. The plurality of augmenters are iteratively retrained using the updated high-precision training set to improve recognition performance over iterations.
    Type: Application
    Filed: June 30, 2023
    Publication date: January 9, 2025
    Inventors: Rakesh Radhakrishnan MENON, Bingqing WANG, Jun ARAKI, Zhengyu ZHOU, Zhe FENG
  • Patent number: 12125478
    Abstract: A computer-implemented method includes receiving one or more word embedding vectors in response to data indicative of one or more words. The method also includes utilizing a first recurrent neural network, outputting one or more intent representation vectors utilizing the one or more word embedding vectors; utilizing a second recurrent neural network, outputting one or more slot representation vectors utilizing the one or more intent representation vectors and the one or more word embedding vectors; and utilizing at least an additional third recurrent neural network and a fourth recurrent neural network, outputting a sentence intent and a slot label based on the one or more word embedding vectors, the one or more intent representation vectors, or the one or more slot representation vectors from either the first or second recurrent neural network.
    Type: Grant
    Filed: August 23, 2021
    Date of Patent: October 22, 2024
    Assignee: Robert Bosch GmbH
    Inventor: Zhengyu Zhou
  • Publication number: 20240231605
    Abstract: A system including a user interface that includes a processor in communication with a display and an input interface, the processor programmed to output on the display the user interface including a keyboard layout, wherein the keyboard layout includes at least a keyboard includes a collection of characters, in response to a first input from the input interface, output a first portion of the keyboard layout associated with a first subset of characters of the keyboard layout, wherein the first subset does not include all of the characters, in response to a second input from the input interface, select a second subset of characters, wherein the second subset of characters is from and include less characters than the first subset of characters and the second subset includes two or more characters, and output a character on a text field associated with the user interface based on the selection of the second subset.
    Type: Application
    Filed: October 25, 2022
    Publication date: July 11, 2024
    Inventors: Jiajing GUO, Nan TIAN, Zhengyu ZHOU, William MA, Nicholas FEFFER, Marcellino GEMELLI
  • Publication number: 20240231580
    Abstract: A virtual reality apparatus that includes a display configured to output information related to a user interface of the virtual reality device, a microphone configured to receive one or more spoken word commands from a user upon activation of a voice recognition session, an eye gaze sensor configured to track eye movement of the user, and a processor programmed to, in response to a first input, output one or more words of a text field, in response to an eye gaze of the user exceeding a threshold time, emphasize a group of one or more words of the text field, toggle through a plurality of words of only the group utilizing the input interface, in response to a second input, highlight and edit an edited word from the group, and in response to utilizing contextual information associated with the group a language model, outputting one or more suggested words.
    Type: Application
    Filed: October 25, 2022
    Publication date: July 11, 2024
    Inventors: Zhengyu ZHOU, Jiajing GUO, Nan TIAN, Nicholas FEFFER, William MA
  • Patent number: 12026366
    Abstract: A system including a user interface that includes a processor in communication with a display and an input interface, the processor programmed to output on the display the user interface including a keyboard layout, wherein the keyboard layout includes at least a keyboard includes a collection of characters, in response to a first input from the input interface, output a first portion of the keyboard layout associated with a first subset of characters of the keyboard layout, wherein the first subset does not include all of the characters, in response to a second input from the input interface, select a second subset of characters, wherein the second subset of characters is from and include less characters than the first subset of characters and the second subset includes two or more characters, and output a character on a text field associated with the user interface based on the selection of the second subset.
    Type: Grant
    Filed: October 25, 2022
    Date of Patent: July 2, 2024
    Assignee: Robert Bosch GmbH
    Inventors: Jiajing Guo, Nan Tian, Zhengyu Zhou, William Ma, Nicholas Feffer, Marcellino Gemelli
  • Publication number: 20240134505
    Abstract: A virtual reality apparatus that includes a display configured to output information related to a user interface of the virtual reality device, a microphone configured to receive one or more spoken word commands from a user upon activation of a voice recognition session, an eye gaze sensor configured to track eye movement of the user, and a processor programmed to, in response to a first input, output one or more words of a text field, in response to an eye gaze of the user exceeding a threshold time, emphasize a group of one or more words of the text field, toggle through a plurality of words of only the group utilizing the input interface, in response to a second input, highlight and edit an edited word from the group, and in response to utilizing contextual information associated with the group a language model, outputting one or more suggested words.
    Type: Application
    Filed: October 24, 2022
    Publication date: April 25, 2024
    Inventors: Zhengyu ZHOU, Jiajing GUO, Nan TIAN, Nicholas FEFFER, William MA
  • Publication number: 20240134516
    Abstract: A system including a user interface that includes a processor in communication with a display and an input interface, the processor programmed to output on the display the user interface including a keyboard layout, wherein the keyboard layout includes at least a keyboard includes a collection of characters, in response to a first input from the input interface, output a first portion of the keyboard layout associated with a first subset of characters of the keyboard layout, wherein the first subset does not include all of the characters, in response to a second input from the input interface, select a second subset of characters, wherein the second subset of characters is from and include less characters than the first subset of characters and the second subset includes two or more characters, and output a character on a text field associated with the user interface based on the selection of the second subset.
    Type: Application
    Filed: October 24, 2022
    Publication date: April 25, 2024
    Inventors: Jiajing GUO, Nan TIAN, Zhengyu ZHOU, William MA, Nicholas FEFFER, Marcellino GEMELLI
  • Publication number: 20240105170
    Abstract: The systems and methods described herein are related to a new speech understanding system for domain-specific voice interaction. The systems and methods described herein combine the automatic correction of an automatic speech recognition errors with a natural language understanding model in a way that optimizes the recognition and understanding of a received speech input. The systems and methods described herein may further support out-of-domain detection or domain classification by jointly learning and/or performing automatic speech recognition error correction and domain-related classification. Through the joint learning with automatic speech recognition error correction, the out-of-domain detection or domain classification may be conducted based on a plurality of possible speech recognition results with shared feature inputs and shared neural layers. The systems and methods described herein may achieve robust performance with high computational efficiency.
    Type: Application
    Filed: September 26, 2022
    Publication date: March 28, 2024
    Inventor: Zhengyu Zhou
  • Patent number: 11848024
    Abstract: A smart mask includes a main body having a back frame and a front cover. The back frame and the front cover each include an opening that is aligned with the mask wearer's mouth when worn. The front cover and back frame may be detachable from one another, or a single piece. A microphone is provided in the main body, as well as a speaker. A processor located in the main body is connected to the microphone and the speaker, and is configured to enhance the speech of the mask wearer. In particular, the processor receives audio signals representing a transformation of a spoken utterance of the wearer, processes the audio signals to enhance the speech, and then outputs the enhanced speech to the speaker. This helps other people better understand what the mask wearer is saying.
    Type: Grant
    Filed: January 26, 2021
    Date of Patent: December 19, 2023
    Assignee: Robert Bosch GmbH
    Inventors: Pongtep Angkititrakul, Xiaoyang Gao, Hyeongsik Kim, Xiaowei Zhou, Zhengyu Zhou
  • Publication number: 20230367966
    Abstract: A method for optimizing performance of a natural language understanding model is disclosed. The method comprises receiving a plurality of training sentences, each training sentence being labeled with a respective intent from a plurality of intents, at least one portion of at least some training sentences being labeled with a respective slot type from a plurality of slot types. The method comprises determining a recommended modification to the plurality of training sentences, the recommended modification including at least one of (i) relabeling and (ii) deleting a training sentence in the plurality of training sentences. The method comprises generating a modified plurality of training sentences by performing the recommended modification to the plurality of training sentences in response to a user input received via a user interface. The method comprises training a first natural language understanding model using the modified plurality of training sentences.
    Type: Application
    Filed: May 11, 2022
    Publication date: November 16, 2023
    Inventor: Zhengyu Zhou
  • Patent number: 11615785
    Abstract: A framework ranks multiple hypotheses generated by one or more ASR engines for each input speech utterance. The framework jointly implements ASR improvement and NLU. It makes use of NLU related knowledge to facilitate the ranking of competing hypotheses, and outputs the top-ranked hypothesis as the improved ASR result together with the NLU results of the speech utterance. The NLU results include intent detection results and the slot filling results.
    Type: Grant
    Filed: May 5, 2020
    Date of Patent: March 28, 2023
    Inventors: Zhengyu Zhou, Xuchen Song
  • Publication number: 20230069049
    Abstract: A computer-implemented method includes receiving one or more word embedding vectors in response to data indicative of one or more words. The method also includes utilizing a first recurrent neural network, outputting one or more intent representation vectors utilizing the one or more word embedding vectors; utilizing a second recurrent neural network, outputting one or more slot representation vectors utilizing the one or more intent representation vectors and the one or more word embedding vectors; and utilizing at least an additional third recurrent neural network and a fourth recurrent neural network, outputting a sentence intent and a slot label based on the one or more word embedding vectors, the one or more intent representation vectors, or the one or more slot representation vectors from either the first or second recurrent neural network.
    Type: Application
    Filed: August 23, 2021
    Publication date: March 2, 2023
    Inventor: Zhengyu ZHOU
  • Publication number: 20220238129
    Abstract: A smart mask includes a main body having a back frame and a front cover. The back frame and the front cover each include an opening that is aligned with the mask wearer's mouth when worn. The front cover and back frame may be detachable from one another, or a single piece. A microphone is provided in the main body, as well as a speaker. A processor located in the main body is connected to the microphone and the speaker, and is configured to enhance the speech of the mask wearer. In particular, the processor receives audio signals representing a transformation of a spoken utterance of the wearer, processes the audio signals to enhance the speech, and then outputs the enhanced speech to the speaker. This helps other people better understand what the mask wearer is saying.
    Type: Application
    Filed: January 26, 2021
    Publication date: July 28, 2022
    Inventors: Pongtep ANGKITITRAKUL, Xiaoyang GAO, Hyeongsik KIM, Xiaowei ZHOU, Zhengyu ZHOU
  • Patent number: 11250853
    Abstract: A dialog system and a method of using the dialog system is disclosed. The method may comprise: receiving audible human speech from a user; determining that the audible human speech comprises sarcasm information; providing an input to a neural network, wherein the input comprises speech data input associated with the audible human speech, an embedding vector associated with the sarcasm information, and a one-hot vector; and based on the input, determining an audible response to the human speech.
    Type: Grant
    Filed: April 30, 2020
    Date of Patent: February 15, 2022
    Assignee: ROBERT BOSCH GMBH
    Inventors: Zhengyu Zhou, In Gyu Choi
  • Publication number: 20210343280
    Abstract: A dialog system and a method of using the dialog system is disclosed. The method may comprise: receiving audible human speech from a user; determining that the audible human speech comprises sarcasm information; providing an input to a neural network, wherein the input comprises speech data input associated with the audible human speech, an embedding vector associated with the sarcasm information, and a one-hot vector; and based on the input, determining an audible response to the human speech.
    Type: Application
    Filed: April 30, 2020
    Publication date: November 4, 2021
    Inventors: Zhengyu ZHOU, In Gyu CHOI
  • Publication number: 20210343288
    Abstract: A spoken dialog system and methods of using the system is described. A method may comprise: receiving audible human speech from a user; determining textual speech data based on the audible human speech; extracting, from the audible human speech, signal speech data that is indicative of acoustic characteristics which correspond to the textual speech data; and using the textual speech data and the signal speech data, generating a response to the audible human speech.
    Type: Application
    Filed: April 30, 2020
    Publication date: November 4, 2021
    Inventors: Zhengyu ZHOU, Vikas YADAV, Yongliang HE, In Gyu CHOI