Patents by Inventor Xiaochun Li

Xiaochun Li 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: 10858211
    Abstract: Provided is a double-channel paper feeding mechanism and a sheet medium processing device. The double-channel paper feeding mechanism includes a fixed frame, a movable frame and a connection mechanism. The fixed frame is provided with a first channel plate and a second channel plate. A first portion of the first channel plate and a first portion of the second channel plate form a conveying channel. An opening for inserting the movable frame is formed between second portions of the two channel plates. The movable frame is provided with a third channel plate and a fourth channel plate. When the movable frame is mounted on the fixed frame, the second portion of the first channel plate and the third channel plate are arranged oppositely and form a first input channel in connection with the conveying channel, the second portion of the second channel plate and the fourth channel plate form a second input channel in connection with the conveying channel.
    Type: Grant
    Filed: March 3, 2017
    Date of Patent: December 8, 2020
    Assignee: Shandong New Beiyang Information Technology Co., Ltd.
    Inventors: Zhenhua Song, Zhengmin Zhang, Ning Feng, Xiaochun Li
  • Publication number: 20200380263
    Abstract: A system for detecting key frames in a video may include a feature extractor configured to extract feature descriptors for each of the multiple image frames in the video. The feature extractor may be an embedded cellular neural network of an artificial intelligence (AI) chip. The system may also include a key frame extractor configured to determine one or more key frames in the multiple image frames based on the corresponding feature descriptors of the image frames. The key frame extractor may determine the key frames based on distance values between a first set of feature descriptors corresponding to a first subset of image frames and a second set of feature descriptors corresponding to a second subset of image frames. The system may output an alert based on determining the key frames and/or display the key frames. The system may also compress the video by removing the non-key frames.
    Type: Application
    Filed: May 29, 2019
    Publication date: December 3, 2020
    Applicant: Gyrfalcon Technology Inc.
    Inventors: Lin Yang, Bin Yang, Qi Dong, Xiaochun Li, Wenhan Zhang, Yinbo Shi, Yequn Zhang
  • Publication number: 20200302276
    Abstract: An artificial intelligence (AI) semiconductor having an embedded convolution neural network (CNN) may include a first convolution layer and a second convolution layer, in which the weights of the first layer and the weights of the second layer are quantized in different bit-widths, thus at different compression ratios. In a VGG neural network, the weights of a first group of convolution layers may have a different compression ratio than the weights of a second group of convolution layers. The weights of the CNN may be obtained in a training system including convolution quantization and/or activation quantization. Depending on the compression ratio, the weights of a convolution layer may be trained with or without re-training. An AI task, such as image retrieval, may be implemented in the AI semiconductor having the CNN described above.
    Type: Application
    Filed: September 27, 2019
    Publication date: September 24, 2020
    Applicant: Gyrfalcon Technology Inc.
    Inventors: Lin Yang, Bin Yang, Hua Zhou, Xiaochun Li, Wenhan Zhang, Qi Dong, Yequn Zhang, Yongxiong Ren, Patrick Dong
  • Publication number: 20200302288
    Abstract: A system for training an artificial intelligence (AI) model for an AI chip may include an AI training unit to train weights of an AI model in floating point, and one or more quantization units for updating the weights of the AI model while accounting for the hardware constraints in the AI chip. The system may also include customization unit for performing one or more linear transformations on the updated weights. The system may also perform output equalization for one or more convolution layers of the AI model to equalize the inputs and/or outputs of each layer of the AI model to within the range allowed in the physical AI chip. The system may further update the weights by performing shift-based quantization that mimics the characteristics of a hardware chip. The updated weights may be stored in fixed point and uploadable to an AI chip implementing an AI task.
    Type: Application
    Filed: September 27, 2019
    Publication date: September 24, 2020
    Applicant: Gyrfalcon Technology Inc.
    Inventors: Yongxiong Ren, Yi Fan, Yequn Zhang, Tianran Chen, Yinbo Shi, Xiaochun Li, Lin Yang
  • Publication number: 20200302289
    Abstract: A system for training an artificial intelligence (AI) model for an AI chip to implement an AI task may include an AI training unit to train weights of an AI model in floating point, a convolution quantization unit for quantizing the trained weights to a number of quantization levels, and an activation quantization unit for updating the weights of the AI model so that output of the AI model based at least on the updated weights are within a range of activation layers of the AI chip. The updated weights may be stored in fixed point and uploadable to the AI chip. The various units may be configured to account for the hardware constraints in the AI chip to minimize performance degradation when the trained weights are uploaded to the AI chip and expedite training convergence. Forward propagation and backward propagation may be combined in training the AI model.
    Type: Application
    Filed: September 27, 2019
    Publication date: September 24, 2020
    Applicant: Gyrfalcon Technology Inc.
    Inventors: Yongxiong Ren, Yi Fan, Yequn Zhang, Baohua Sun, Bin Yang, Xiaochun Li, Lin Yang
  • Publication number: 20200293865
    Abstract: A cellular neural network architecture may include a processor and embedded cellular, neural network (CeNN) executable in an artificial intelligence (AI) integrated circuit and configured to perform certain AI functions. The CeNN may include multiple convolution layers, each having multiple binary weights. In some examples, a method may configure a given layer of the CeNN and one or more additional layers of the CeNN to retrieve the output of the given layer for debugging or training the CeNN. In configuring the one or more additional layers, the method may use an identity layer.
    Type: Application
    Filed: March 14, 2019
    Publication date: September 17, 2020
    Applicant: Gyrfalcon Technology Inc.
    Inventors: Bowei Liu, Yinbo Shi, Yequn Zhang, Xiaochun Li
  • Publication number: 20200293856
    Abstract: A cellular neural network architecture may include a processor and an embedded cellular neural network (CeNN) executable in an artificial intelligence (AI) integrated circuit and configured to perform certain AI functions. The CeNN may include multiple convolution layers, such as first, second, and third layers, each layer having multiple binary weights. In some examples, a method may configure the multiple layers in the CeNN to produce a residual connection. In configuring the second and third layers, the method may use an identity matrix.
    Type: Application
    Filed: March 14, 2019
    Publication date: September 17, 2020
    Applicant: Gyrfalcon Technology Inc.
    Inventors: Bowei Liu, Yinbo Shi, Yequn Zhang, Xiaochun Li
  • Publication number: 20200250523
    Abstract: In some examples, given an AI model in floating point, a system may use one or more artificial intelligence (AI) chips to train a global gain vector for use to convert the AI model in floating point to an AI model in fixed point for uploading to a physical AI chip. The system may determine initial gain vectors, and in each of multiple iterations, obtain the performance values of the AI chips based on the gain vectors and update the gam vectors for the next iteration. The gain vectors are updated based on a velocity of gain. The performance value may be based on feature maps of an AI model before and after the converting. The performance value may also be based on interference over a test dataset. Upon completion of the iterations, the system determines the global gain vector that resulted in the best performance value during the iterations.
    Type: Application
    Filed: February 5, 2019
    Publication date: August 6, 2020
    Applicant: Gyrfalcon Technology Inc.
    Inventors: Yongxiong Ren, Yequn Zhang, Baohua Sun, Xiaochun Li, Qi Dong, Lin Yang
  • Publication number: 20200244963
    Abstract: A sample characterization system is disclosed. In embodiments, the sample characterization system includes a controller communicatively coupled to an inspection sub-system, the controller including one or more processors configured to execute a set of program instructions stored in memory, the set of program instructions configured to cause the one or more processors to: acquire one or more target image frames of a sample; generate a target tensor with the one or more acquired target image frames; perform a first set of one or more decomposition processes on the target tensor to form generate one or more reference tensors including one or more reference image frames; identify one or more differences between the one or more target image frames and the one or more reference image frames; and determine one or more characteristics of the sample based on the one or more identified differences.
    Type: Application
    Filed: January 16, 2020
    Publication date: July 30, 2020
    Inventors: Nurmohammed Patwary, Richard Wallingford, James A. Smith, Xiaochun Li, Vladimir Tumakov, Bjorn Brauer
  • Publication number: 20200234118
    Abstract: A system may include multiple client devices and a processing device communicatively coupled to the client devices. One or more client devices may implement a greedy approach in searching for an optimal artificial intelligence (AI) model. For example, a client device may use a training dataset to perform an AI task, and update its AI model. The client device may verify the performance of the AI task and determine whether to accept or reject its updated AI model. Upon rejection, the client device may repeat updating its AI model until the updated AI model is accepted, or until a stopping criteria is met. The processing device may be configured to update the initial AI models based on the accepted updated AI models obtained in the multiple client device. Training data for each of the client devices may contain a subset shuffled from a larger training dataset.
    Type: Application
    Filed: December 3, 2019
    Publication date: July 23, 2020
    Applicant: Gyrfalcon Technology Inc.
    Inventors: Yinbo Shi, Yequn Zhang, Xiaochun Li, Bowei Liu
  • Publication number: 20200234119
    Abstract: A system may include multiple client devices and a processing device communicatively coupled to the client devices. A client device may receive an initial artificial intelligence (AI) model, use a training dataset to perform an AI task, and update its AI model. The client device may verify the performance of the AI task to determine whether to accept or reject its updated AI model. Upon rejection, the client device may repeat updating its AI model until the updated AI model is accepted, or until a stopping criteria is met. The processing device may be configured to update the initial AI models based on the accepted updated AI models obtained in the multiple client devices, and repeat the process for each client device using the updated initial AI models. Training data for each of the client devices may contain a subset shuffled from a larger training dataset.
    Type: Application
    Filed: December 3, 2019
    Publication date: July 23, 2020
    Applicant: Gyrfalcon Technology Inc.
    Inventors: Yinbo Shi, Yequn Zhang, Xiaochun Li, Bowei Liu
  • Publication number: 20200151558
    Abstract: A system may be configured to obtain a global artificial intelligence (AI) model for uploading into an AI chip to perform AI tasks. The system may implement a training process including receiving updated AI models from one or more client devices, determining a global AI model based on the received AI models from the client devices, and updating initial AI models for the client devices. Each client device may receive an initial AI model and train an updated AI model by training the entire parameters of the AI model together, by training a subset of the parameters of the AI model in a layer by layer fashion, or by training a subset of the parameters by parameter types. Each client device may include one or more AI chips configured to run an AI task to measure performance of an AI model. The AI model may include a convolutional neural network.
    Type: Application
    Filed: February 11, 2019
    Publication date: May 14, 2020
    Applicant: Gyrfalcon Technology Inc.
    Inventors: Yongxiong Ren, Yequn Zhang, Baohua Sun, Xiaochun Li, Qi Dong, Lin Yang
  • Patent number: 10640850
    Abstract: Solid immiscible alloys and methods for making the solid immiscible alloys are provided. The microstructure of the immiscible alloys is characterized by a minority phase comprising a plurality of particles of an inorganic material dispersed in a majority phase comprising a continuous matrix of another inorganic material. The methods utilize nanoparticles to control both the collisional growth and the diffusional growth of the minority phase particles in the matrix during the formation of the alloy microstructure.
    Type: Grant
    Filed: June 27, 2018
    Date of Patent: May 5, 2020
    Assignee: Wisconsin Alumni Research Foundation
    Inventors: Xiaochun Li, Lianyi Chen, Jiaquan Xu
  • Publication number: 20200126830
    Abstract: A method of semiconductor-wafer image alignment is performed at a semiconductor-wafer defect-inspection system. In the method, a semiconductor wafer is loaded into the semiconductor-wafer defect-inspection system. Pre-inspection alignment is performed for the semiconductor wafer. After performing the pre-inspection alignment, a first swath is executed to generate a first image of a first region on the semiconductor wafer. An offset of a target structure in the first image with respect to a known point is determined. Defect identification is performed for the first image, using the offset. After executing the first swath and determining the offset, a second swath is executed to generate a second image of a second region on the semiconductor wafer. While executing the second swath, run-time alignment of the semiconductor wafer is performed using the offset.
    Type: Application
    Filed: May 21, 2019
    Publication date: April 23, 2020
    Inventors: David Dowling, Tarunark Singh, Bjorn Brauer, Santosh Bhattacharyya, Bryant Mantiply, Hucheng Lee, Xiaochun Li, Sangbong Park
  • Patent number: 10513759
    Abstract: A manufacturing method includes: 1) forming a melt including one or more metals; 2) introducing nanostructures into the melt at an initial volume fraction of the nanostructures; and 3) at least partially evaporating one or more metals from the melt so as to form a metal matrix nanocomposite including the nanostructures dispersed therein at a higher volume fraction than the initial volume fraction.
    Type: Grant
    Filed: January 19, 2017
    Date of Patent: December 24, 2019
    Assignee: The Regents of the University of California
    Inventors: Xiaochun Li, Lianyi Chen
  • Publication number: 20190313833
    Abstract: A rice cooker assembly uses machine learning models to identify and classify different types of food stored. The rice cooker has a chamber including different compartments for storing different types of food. A camera is positioned to view an interior of the chamber. The camera captures images of the contents of the chamber. From the images, the machine learning model classifies the different types of food stored. The rice cooker determines a mixture of different types of food based on nutrition value and/or taste. The rice cooker creates the mixture and controls the cooking process accordingly. The one or more machine learning models may be resident in the rice cooker or it may be accessed via a network.
    Type: Application
    Filed: April 16, 2018
    Publication date: October 17, 2019
    Inventors: Xiaochun Li, Linnan Zhu, Hua Zhou
  • Publication number: 20190291213
    Abstract: Surface asperities, such as roughness characteristics, are reduced or otherwise mitigated via the control of surface regions including the asperities in different regimes. In accordance with various embodiments, the height of both high-frequency and low-frequency surface asperities is reduced by controlling characteristics of a surface region under a first regime to flow material from the surface asperities. A second regime is implemented to reduce a height of high-frequency surface asperities in the surface region by controlling characteristics of the surface region under a second regime to flow material that is predominantly from the high-frequency surface asperities, the controlled characteristics in the second regime being different than the controlled characteristics in the first regime. Such aspects may include, for example, controlling melt pools in each regime via energy pulses, to respectively mitigate/reduce the asperities.
    Type: Application
    Filed: June 10, 2019
    Publication date: September 26, 2019
    Inventors: Venkata Madhukanth Vadali, Chao Ma, Neil Arthur Duffie, Xiaochun Li, Frank Ewald Pfefferkorn
  • Publication number: 20190274337
    Abstract: A rice cooker assembly uses machine learning models to identify and classify content in grain mixtures thereby to provide better automation of the cooking process. As one example, a rice cooker has a chamber storing grains. A camera is positioned to view an interior of the chamber. The camera captures images of the contents of the chamber. From the images, the machine learning model determines whether the contents of the chamber includes one type or multiple types of grain or whether the contents of the chamber includes any inedible objects. The machine learning model further classifies the one or more types of grains and inedible objects if any. The cooking process may be controlled accordingly. The machine learning model may be resident in the rice cooker or it may be accessed via a network.
    Type: Application
    Filed: March 8, 2018
    Publication date: September 12, 2019
    Inventors: Dongyan Wang, Linnan Zhu, Xiaochun Li, Junyang Zhou
  • Patent number: D901123
    Type: Grant
    Filed: June 10, 2019
    Date of Patent: November 3, 2020
    Assignee: GUANGDONG WIREKING COMMERCIAL CO., LTD.
    Inventors: Minyu Liu, Xiaochun Li
  • Patent number: D901817
    Type: Grant
    Filed: May 30, 2019
    Date of Patent: November 10, 2020
    Assignee: GUANGDONG WIREKING HOUSEWARES & HARDWARE CO., LTD
    Inventors: Xiaochun Li, Weitao Chen