Patents by Inventor Gary Wang
Gary Wang 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: 20250105249Abstract: The present disclosure provides a method for producing a LGPS-based solid electrolyte at a significantly lower temperature and/or time compared to conventional methods. In one particular embodiment, methods disclosed herein include subjecting a material comprising an LGPS-based powder to a pressure of at least 250 MPa at a temperature less than about 450° C. under conditions sufficient to produce the LGPS-based solid electrolyte.Type: ApplicationFiled: January 21, 2022Publication date: March 27, 2025Inventors: Alexander KOZEN, Sang Bok LEE, Gary W. RUBLOFF, Cholho LEE, Yang WANG
-
Publication number: 20250095639Abstract: A method includes receiving a set of training utterances each including a non-synthetic speech representation of a corresponding utterance, and for each training utterance, generating a corresponding synthetic speech representation by using a voice conversion model. The non-synthetic speech representation and the synthetic speech representation form a corresponding training utterance pair. At each of a plurality of output steps for each training utterance pair, the method also includes generating, for output by a speech recognition model, a first probability distribution over possible non-synthetic speech recognition hypotheses for the non-synthetic speech representation and a second probability distribution over possible synthetic speech recognition hypotheses for the synthetic speech representation.Type: ApplicationFiled: November 27, 2024Publication date: March 20, 2025Applicant: Google LLCInventors: Andrew M. Rosenberg, Gary Wang, Bhuvana Ramabhadran, Fadi Biadsy
-
Publication number: 20250094805Abstract: A method and system for polymorphic pruning of neural networks are disclosed. The method involves training a neural network model on an input dataset for an initial predetermined number of iterations to gather weight information, including the strength of each weight and changes in strength over iterations. The weights are stored in an array accessible to a pruning algorithm. An objective function is compiled using the weight information, and an optimization tool solves the objective function to generate a solution vector. This solution vector is used to create a pruning mask, which is applied to the neural network model to prune certain weights by setting them to zero. The pruned weight vector updates the model, resulting in a neural network with fewer non-zero connections between neurons.Type: ApplicationFiled: September 19, 2024Publication date: March 20, 2025Applicant: Entanglement, Inc.Inventors: Anna HUGHES, Amit VERMA, Haibo WANG, Gary KOCHENBERGER, Fred GLOVER, Amit HULANDAGERI
-
Patent number: 12253668Abstract: Embodiments of the disclosure provide a scanning mirror assembly. In certain configurations, the scanning mirror assembly may include a two-dimensional micro-electromechanical system (MEMS) scanning mirror, a first pair of piezoelectric electrodes coupled to the MEMS scanning mirror through a first pair of looped torsion springs, and a second pair of piezoelectric electrodes coupled to the MEMS scanning mirror through a second pair of looped torsion springs. The first pair of piezoelectric electrodes drives the MEMS scanning mirror to rotate around a first axis. The second pair of piezoelectric electrodes drives the MEMS scanning mirror to rotate around a second axis orthogonal to the first axis.Type: GrantFiled: December 30, 2021Date of Patent: March 18, 2025Assignee: BEIJING VOYAGER TECHNOLOGY CO., LTD.Inventors: Youmin Wang, Yufeng Wang, Kong Yin Ho, Gary Li
-
Patent number: 12243496Abstract: Often when there is a glare on a display screen the user may be able to mitigate the glare by tilting or otherwise moving the screen or changing their viewing position. However, when driving a car there are limited options for overcoming glares on the dashboard, especially when you are driving for a long distance in the same direction. Embodiments are directed to eliminating such glare. Other embodiments are related to mixed reality (MR) and filling in occluded areas.Type: GrantFiled: May 24, 2023Date of Patent: March 4, 2025Assignee: Intel CorporationInventors: Arthur J. Runyan, Richmond Hicks, Nausheen Ansari, Narayan Biswal, Ya-Ti Peng, Abhishek R. Appu, Wen-Fu Kao, Sang-Hee Lee, Joydeep Ray, Changliang Wang, Satyanarayana Avadhanam, Scott Janus, Gary Smith, Nilesh V. Shah, Keith W. Rowe, Robert J. Johnston
-
Patent number: 12236094Abstract: The subject disclosure relates to storage. An example method includes receiving an input/output (I/O) request based on a protocol related to a first type of storage medium. The method further includes determining a namespace among a plurality of namespaces that is associated with the I/O request. The method further includes determining a binary file corresponding to the determined namespace based on a mapping relationship. The method further includes reading from and/or writing to a second type of storage medium based on the determined binary file, the second type being different from the first type. In this manner, a unified underlying storage implementation with software definition capabilities is provided, which enables different types of storage media to be operated as namespaces, thereby enabling the deployment of a storage system to adapt to more platforms and scenarios.Type: GrantFiled: August 30, 2023Date of Patent: February 25, 2025Assignee: DELL PRODUCTS L.P.Inventors: Xingshan Wang, Haiyan He, Gary Jialei Wu
-
Publication number: 20250054767Abstract: Embodiments include a method of forming a contact structure on a semiconductor substrate. The method including selectively depositing a metal silicide layer over a contact formed within a cavity of a substrate and a bottom surface of the cavity using a selective deposition process, including forming a residual layer on a surface of a dielectric layer forming sidewalls of the cavity, wherein a thickness of the metal silicide layer deposited over the contact is greater than a thickness of the residual layer, removing at least a portion of the residual layer formed on the dielectric layer using an etching process that comprises exposing the metal selectively deposited layer to a metal halide containing precursor, and selectively depositing a metal fill over the metal silicide layer remaining over the contact after removing the at least the portion of the residual layer using a selective metal fill process.Type: ApplicationFiled: April 25, 2024Publication date: February 13, 2025Inventors: Qihao ZHU, Shumao ZHANG, Weifeng YE, Yiyang WAN, Gary HOW, Jianqiu GUO, Dong WANG, Shihchung CHEN, Liqi WU, Jiang LU
-
Publication number: 20250054812Abstract: Embodiments include a method of forming a contact structure on a semiconductor substrate. The method including selectively depositing a metal silicide layer over a contact formed within a cavity of a substrate and a bottom surface of the cavity using a selective deposition process, including forming a residual layer on a surface of a dielectric layer forming sidewalls of the cavity, wherein a thickness of the metal silicide layer deposited over the contact is greater than a thickness of the residual layer, removing at least a portion of the residual layer formed on the dielectric layer using an etching process that comprises exposing the metal selectively deposited layer to a metal halide containing precursor, and selectively depositing a metal fill over the metal silicide layer remaining over the contact after removing the at least the portion of the residual layer using a selective metal fill process.Type: ApplicationFiled: December 29, 2023Publication date: February 13, 2025Inventors: Qihao ZHU, Shumao ZHANG, Weifeng YE, Yiyang WAN, Gary HOW, Jianqiu GUO, Dong WANG, Shihchung CHEN, Liqi WU, Jiang LU
-
Publication number: 20250046466Abstract: Techniques for predicting analyte levels based on intermittent sensor data are disclosed. First sensor data is acquired from an analyte sensor. The first sensor data reflects analyte levels of a user who is wearing the analyte sensor. The first sensor data is collected over a first time period. Later, a determination is made as to whether data is still being acquired from the analyte sensor. As a result of determining that data is no longer being acquired from the analyte sensor, the first sensor data is classified as intermittent analyte data. The intermittent analyte data is then used to generate predicted analyte level data for the user during a second time period that is subsequent to the first time period. The predicted analyte level data is reflective of the intermittent analyte data.Type: ApplicationFiled: July 29, 2024Publication date: February 6, 2025Inventors: Shridhara Alva KARINKA, Vivek S. RAO, Chung-Che WANG, Peter Gary ROBINSON, Jennifer Lynn OLSON, Xuandong HUA, Panganamala Ashwin KUMAR, Sujit R. JANGAM
-
Patent number: 12208585Abstract: Methods of manufacturing composite workpieces that include adding an expandable element to an internal volume of a constraining container proximate to a uncured composite workpiece supported on a rigid form, where the expandable element is configured to expand when a predetermined change is produced in an attribute of the expandable element; expanding the expandable element by producing the predetermined change in the attribute of the expandable element, so that an expansion of the expandable element applies pressure to the workpiece supported on the rigid form within the internal volume, and curing the composite workpiece while the resulting pressure is applied to the workpiece supported on the rigid form.Type: GrantFiled: October 24, 2023Date of Patent: January 28, 2025Assignee: The Boeing CompanyInventors: Jonathan A. Santiago, Xiaoxi Wang, Gary Ernest Georgeson
-
Publication number: 20250022691Abstract: A wide-coverage edge ring configured to be arranged above a bottom ring in a substrate processing chamber includes an upper surface, a lower surface that includes a lower surface step that extends downward from the lower surface and is configured to be received within and interface with a pocket defined at least partially by an upper surface of the bottom ring and an inner surface of a chamber liner, an inner diameter, a ledge defined in the inner diameter of the edge ring, and an outer diameter. The outer diameter of the edge ring includes a projection that extends radially outward from the edge ring and defines an inward step in the outer diameter, the projection and the inward step are configured to interface with an upper end of the chamber liner, and the projection is configured to extend at least partially over the upper end of the chamber liner.Type: ApplicationFiled: February 7, 2022Publication date: January 16, 2025Inventors: Shang-I CHOU, Dandan WANG, Giovanni Paolo GAJUDO, Gary GRAGG, Robert O'NEILL
-
Patent number: 12190862Abstract: A method includes receiving a set of training utterances each including a non-synthetic speech representation of a corresponding utterance, and for each training utterance, generating a corresponding synthetic speech representation by using a voice conversion model. The non-synthetic speech representation and the synthetic speech representation form a corresponding training utterance pair. At each of a plurality of output steps for each training utterance pair, the method also includes generating, for output by a speech recognition model, a first probability distribution over possible non-synthetic speech recognition hypotheses for the non-synthetic speech representation and a second probability distribution over possible synthetic speech recognition hypotheses for the synthetic speech representation.Type: GrantFiled: April 25, 2022Date of Patent: January 7, 2025Assignee: Google LLCInventors: Andrew M. Rosenberg, Gary Wang, Bhuvana Ramabhadran, Fadi Biadsy
-
Publication number: 20240304178Abstract: A method includes receiving training data including transcribed speech utterances spoken in a general domain, modified speech utterances in a target domain, and unspoken textual utterances corresponding to the transcriptions of the modified speech utterances in the target domain. The modified speech utterances include utterances spoken in the target domain that have been modified to obfuscate one or more classes of sensitive information recited in the utterances. The method also includes generating a corresponding alignment output for each unspoken textual utterance of the received training data using an alignment model. The method also includes training a speech recognition model on the alignment outputs generated for the corresponding to the unspoken textual utterances, the un-transcribed speech utterances, and the transcribed speech utterances to teach the speech recognition model to learn to recognize speech in the target domain and phrases within the one or more classes of sensitive information.Type: ApplicationFiled: February 12, 2024Publication date: September 12, 2024Applicant: Google LLCInventors: Andrew M Rosenberg, Yacob Yochai Blau, Bhuvana Ramabhadran, Genady Beryozkin, Gary Wang, Zhehuai Chen, Rohan Agrawal, Parisa Haghani
-
Patent number: 12041983Abstract: Generally described, aspects of the disclosed subject matter are directed to impact mitigating membranes. In accordance with aspects of the present disclosure, the impact mitigating membranes generally include a laminar core having a first lamina and a second lamina coupled to the first lamina and configured to shift relative to the first lamina to dissipate kinetic energy resulting from an oblique impact force. The laminar core may include a lubricating layer or self-lubrication between the lamina to reduce friction. In general, the laminar core is surrounded by first and second encapsulating layers positioned adjacent each lamina. An anchoring attachment coupled to one or more of the first and second encapsulating layer and associating with the laminar core constrains relative lateral motion between the laminar core and the first and second encapsulating layers.Type: GrantFiled: March 6, 2018Date of Patent: July 23, 2024Assignee: Simon Fraser UniversityInventors: Daniel Eamon Abram, Farid Golnaraghi, Gaofeng Gary Wang
-
Publication number: 20230298565Abstract: A method includes receiving a set of training utterances each including a non-synthetic speech representation of a corresponding utterance, and for each training utterance, generating a corresponding synthetic speech representation by using a voice conversion model. The non-synthetic speech representation and the synthetic speech representation form a corresponding training utterance pair. At each of a plurality of output steps for each training utterance pair, the method also includes generating, for output by a speech recognition model, a first probability distribution over possible non-synthetic speech recognition hypotheses for the non-synthetic speech representation and a second probability distribution over possible synthetic speech recognition hypotheses for the synthetic speech representation.Type: ApplicationFiled: April 25, 2022Publication date: September 21, 2023Applicant: Google LLCInventors: Andrew M. Rosenberg, Gary Wang, Bhuvana Ramabhadran, Fadi Biadsy
-
Patent number: 11684095Abstract: The present disclosure seeks to reduce the effects of rotational and linear acceleration experienced by the body of a user in response to an impact force. Modular disengaging systems of the present disclosure are generally suitable for coupling to protective equipment to provide a disengaging motion between various layers such that the effects of the impact force to the body of the user are reduced. Generally described, the modular disengaging systems of the present disclosure include layers configured to facilitate relative lateral motion therebetween upon an impact force.Type: GrantFiled: September 21, 2018Date of Patent: June 27, 2023Assignee: Simon Fraser UniversityInventors: Daniel Eamon Abram, Farid Golnaraghi, Gaofeng Gary Wang, Combiz Jelveh
-
Publication number: 20230013587Abstract: A method includes receiving training data that includes unspoken text utterances, un-transcribed non-synthetic speech utterances, and transcribed non-synthetic speech utterances. Each unspoken text utterance is not paired with any corresponding spoken utterance of non-synthetic speech. Each un-transcribed non-synthetic speech utterance is not paired with a corresponding transcription. Each transcribed non-synthetic speech utterance is paired with a corresponding transcription. The method also includes generating a corresponding synthetic speech representation for each unspoken textual utterance of the received training data using a text-to-speech model. The method also includes pre-training an audio encoder on the synthetic speech representations generated for the unspoken textual utterances, the un-transcribed non-synthetic speech utterances, and the transcribed non-synthetic speech utterances to teach the audio encoder to jointly learn shared speech and text representations.Type: ApplicationFiled: April 15, 2022Publication date: January 19, 2023Applicant: Google LLCInventors: Andrew Rosenberg, Zhehuai Chen, Bhuvana Ramabhadran, Pedro J. Moreno Mengibar, Gary Wang, Yu Zhang
-
Publication number: 20220310065Abstract: A method includes receiving audio data corresponding to an utterance and generating a pair of positive audio data examples. Here, each positive audio data example includes a respective augmented copy of the received audio data. For each respective positive audio data example, the method includes generating a respective sequence of encoder outputs and projecting the respective sequence of encoder outputs for the positive data example into a contrastive loss space. The method also includes determining a L2 distance between each corresponding encoder output in the projected sequences of encoder outputs for the positive audio data examples and determining a per-utterance consistency loss by averaging the L2 distances. The method also includes generating corresponding speech recognition results for each respective positive audio data example. The method also includes updating parameters of the speech recognition model based on a respective supervised loss term and the per-utterance consistency loss.Type: ApplicationFiled: March 22, 2022Publication date: September 29, 2022Applicant: Google LLCInventors: Andrew Rosenberg, Bhuvana Ramabhadran, Zhehuai Chen, Gary Wang, Yu Zhang, Jesse Emond
-
Publication number: 20210371303Abstract: A water filter includes a housing, a spoiler unit, a power generation unit, a sterilization unit and a purification unit. The housing includes an input portion, an output portion and a first thread portion. The spoiler unit is provided in the housing and corresponds to the input portion. The power generation unit is provided in the housing and is coupled to the spoiler unit. The sterilization unit is provided in the housing and is electrically connected to the power generation unit, and includes a sterilization light source. The purification unit is provided in the housing and corresponds to the output portion and the sterilization unit, and includes a plurality of purification particles. Thus, the water filter is enabled to be coupled to a faucet using the first thread portion, allowing tap water to flow toward the direction from the input portion to the output portion.Type: ApplicationFiled: June 2, 2020Publication date: December 2, 2021Inventors: HSIU-LING YANG, GARY WANG
-
Publication number: 20210137184Abstract: Generally described, aspects of the disclosed subject matter are directed to impact mitigating membranes. In accordance with aspects of the present disclosure, the impact mitigating membranes generally include a laminar core having a first lamina and a second lamina coupled to the first lamina and configured to shift relative to the first lamina to dissipate kinetic energy resulting from an oblique impact force. The laminar core may include a lubricating layer or self-lubrication between the lamina to reduce friction. In general, the laminar core is surrounded by first and second encapsulating layers positioned adjacent each lamina. An anchoring attachment coupled to one or more of the first and second encapsulating layer and associating with the laminar core constrains relative lateral motion between the laminar core and the first and second encapsulating layers.Type: ApplicationFiled: March 6, 2018Publication date: May 13, 2021Applicant: SIMON FRASER UNIVERSITYInventors: Daniel Eamon ABRAM, Farid GOLNARAGHI, Gaofeng Gary WANG