Patents Examined by Casey R. Garner
  • Patent number: 12033043
    Abstract: An identification method of urban functional areas based on mixing degree of functions and integrated learning includes the following steps: 1) performing data acquisition and preprocessing; 2) constructing 10 indicator features of an urban functional area identification system; 3) structuring the indicator features: acquiring, by a spatial statistical tool, the 10 indicator features corresponding to each parcel; 4) constructing an independent variable dataset; 5) labeling response variables; 6) dividing a training dataset into a plurality of training subsets according to the mixing degree of functions; 7) training a Stacking-based integrated learning model; and 8) joining an attribute in one table to another table, so as to complete the identification of the urban functional areas on each parcel. The identification method divides the training dataset by grading the mixing degree of functions, and makes predictions based on the prediction dataset with corresponding mixing degree of functions.
    Type: Grant
    Filed: July 1, 2022
    Date of Patent: July 9, 2024
    Assignee: NANJING UNIVERSITY
    Inventors: Chen Zhou, Yunyun Xu, Zhenjie Chen, Manchun Li, Xin Zhao, Zhixin Yu, Haoyang Du, Boqing Wen, Zian Wang, Nan Xia
  • Patent number: 12020129
    Abstract: A technique for training a model includes obtaining a training example for a model having model parameters stored on one or more computer readable storage mediums operably coupled to the hardware processor. The training example includes an outcome and features to explain the outcome. A gradient is calculated with respect to the model parameters of the model using the training example. Two estimates of a moment of the gradient with two different time constants are computed for the same type of the moment using the gradient. Using a hardware processor, the model parameters of the model are updated using the two estimates of the moment with the two different time constants to reduce errors while calculating the at least two estimates of the moment of the gradient.
    Type: Grant
    Filed: April 13, 2023
    Date of Patent: June 25, 2024
    Assignee: INTERNATIONAL BUSINESS MACHINES CORPORATION
    Inventor: Tetsuro Morimura
  • Patent number: 12014254
    Abstract: A machine-learning based method includes receiving an instruction to model an engagement predicting score for a user. User-specific, activity-specific data is obtained from digital resources that include a user-specific activity performance data regarding performance of at least one activity by the user, an object data for an object that allows the user to perform the at least one activity, and user-specific personal data of the user. A user-specific activity engagement labeling data for the at least one activity is predicted by utilizing a first-type data pipeline on the at least one user-specific activity performance data. User-specific, activity-specific data features are predicted by utilizing a second-type data pipeline on the user-specific, activity-specific data. The engagement predicting score is predicted from the user-specific, activity-specific data features and the user-specific activity engagement labeling data.
    Type: Grant
    Filed: December 29, 2022
    Date of Patent: June 18, 2024
    Assignee: Broadridge Financial Solutions, Inc.
    Inventors: Richard Bryce, Joseph Lo, Luca Marchesotti
  • Patent number: 12008478
    Abstract: Systems and methods for training and utilizing constrained generative models in accordance with embodiments of the invention are illustrated. One embodiment includes a method for training a constrained generative model. The method includes steps for receiving a set of data samples from a first distribution, identifying a set of constraints from a second distribution, and training a generative model based on the set of data samples and the set of constraints.
    Type: Grant
    Filed: October 19, 2020
    Date of Patent: June 11, 2024
    Assignee: Unlearn.AI, Inc.
    Inventors: Aaron M. Smith, Anton D. Loukianov, Charles K. Fisher, Jonathan R. Walsh
  • Patent number: 11989653
    Abstract: A system for increasing quality of results of computations of an artificial neural network (ANN) by using complex rounding rules for parameters in the ANN is provided, the system comprising one or more processing units configured to: receive a plurality of first parameters for one or more neurons of ANN, the first parameters being of a first data type; and change the first parameters to second parameters of a second data type to obtain a plurality of the second parameters according to a rule in which a distance between at least one first parameter and corresponding second parameter is greater than a distance between the first parameter and a value of the second data type closest to the at least one first parameter. A distance between a vector of the first parameters and a vector of the second parameters is minimized.
    Type: Grant
    Filed: May 22, 2020
    Date of Patent: May 21, 2024
    Assignee: Mipsology SAS
    Inventor: Gabriel Gouvine
  • Patent number: 11989648
    Abstract: A training log is selected from a plurality of well logs. A log window of a plurality of log windows is selected from the training log. A positive window is generated from the log window. A negative window is selected from the training log. A siamese neural network (SNN) is trained that includes a first self attention neural network (ANN) and a duplicate self attention neural network with the log window, the positive window, and the negative window, to recognize a similarity between the log window and the positive window and to differentiate against the negative window.
    Type: Grant
    Filed: September 11, 2020
    Date of Patent: May 21, 2024
    Assignee: SCHLUMBERGER TECHNOLOGY CORPORATION
    Inventors: Mandar Shrikant Kulkarni, Hiren Maniar, Aria Abubakar
  • Patent number: 11966826
    Abstract: One or more default protected attribute values may be determined for a prediction model trained based on training data including a plurality of training observations. Each of the plurality of training observations may include a respective plurality of training data values corresponding with a plurality of features. Each of the plurality of training observations may also include a respective target value. Each of the plurality of training observations may include a respective protected attribute value corresponding with a protected attribute feature. A request to determine a designated predicted target value for a designated inference observation may be received after determining the one or more default protected attribute values. The predicted target value may be determined by applying the prediction model to an inference observation and a designated default protected attribute value of the one or more default protected attribute values.
    Type: Grant
    Filed: November 14, 2022
    Date of Patent: April 23, 2024
    Assignee: Epistamai LLC
    Inventor: Christopher Lam
  • Patent number: 11954594
    Abstract: This document generally describes a neural network training system, including one or more computers, that trains a recurrent neural network (RNN) to receive an input, e.g., an input sequence, and to generate a sequence of outputs from the input sequence. In some implementations, training can include, for each position after an initial position in a training target sequence, selecting a preceding output of the RNN to provide as input to the RNN at the position, including determining whether to select as the preceding output (i) a true output in a preceding position in the output order or (ii) a value derived from an output of the RNN for the preceding position in an output order generated in accordance with current values of the parameters of the recurrent neural network.
    Type: Grant
    Filed: May 10, 2021
    Date of Patent: April 9, 2024
    Assignee: Google LLC
    Inventors: Samy Bengio, Oriol Vinyals, Navdeep Jaitly, Noam M. Shazeer
  • Patent number: 11941520
    Abstract: Techniques regarding determining hyperparameters for a differentially private federated learning process are provided. For example, one or more embodiments described herein can comprise a system, which can comprise a memory that can store computer executable components. The system can also comprise a processor, operably coupled to the memory, and that can execute the computer executable components stored in the memory. The computer executable components can comprise a hyperparameter advisor component that determines a hyperparameter for a model of a differentially private federated learning process based on a defined numeric relationship between a privacy budget, a learning rate schedule, and a batch size.
    Type: Grant
    Filed: January 9, 2020
    Date of Patent: March 26, 2024
    Assignee: INTERNATIONAL BUSINESS MACHINES CORPORATION
    Inventors: Colin Sutcher-Shepard, Ashish Verma, Jayaram Kallapalayam Radhakrishnan, Gegi Thomas
  • Patent number: 11928607
    Abstract: Embodiments are directed to managing data for a predictive learner recommendation platform. A platform that includes applications hosted in an application layer may be provided. The applications may be employed to provide a request to determine a pathway prediction for a learner such that the pathway prediction may be associated with a role offered by employers. Prediction engines associated with the request may be determined based on the service layer interface and the request such that the request may be provided to the determined prediction engines via the service layer interface. The prediction engines may be employed to generate the pathway prediction based on a learner profile that corresponds with the learner, a role success profile that corresponds to the employers, and models that are trained to predict matches between the learner profile and the role success profile.
    Type: Grant
    Filed: August 16, 2022
    Date of Patent: March 12, 2024
    Assignee: AstrumU, Inc.
    Inventors: Adam Jason Wray, Kaj Orla Peter Pedersen, Xiao Cai, Jue Gong
  • Patent number: 11928590
    Abstract: A device includes a state machine. The state machine includes a plurality of blocks, where each of the blocks includes a plurality of rows. Each of these rows includes a plurality of programmable elements. Furthermore, each of the programmable elements are configured to analyze at least a portion of a data stream and to selectively output a result of the analysis. Each of the plurality of blocks also has corresponding block activation logic configured to dynamically power-up the block.
    Type: Grant
    Filed: January 28, 2021
    Date of Patent: March 12, 2024
    Assignee: Micron Technology, Inc.
    Inventor: Harold B Noyes
  • Patent number: 11922327
    Abstract: Domain specific knowledge base (KB) contains all concepts from domain and the semantic relations between concepts. The concepts and the semantic relations are extracted from an existing corpus of content for the domain. The World Wide Web Consortium (W3C) standard SKOS (Simple Knowledge Organization System) can be used and two types of semantic relations can be captured: hierarchal and associative. A Natural Language Processing (NLP) software engine can parse the input text to create a semantic knowledge graph, which is then mapped to a SKOS knowledge model. During the linguistic understanding of the text, relevant domain concepts are identified and connected by semantic links. Concepts automatically identified as most important in this domain can be promoted to another layer, referred to as “Topics.
    Type: Grant
    Filed: November 28, 2022
    Date of Patent: March 5, 2024
    Assignee: Morgan Stanley Services Group Inc.
    Inventors: Nicolas Seyot, Richard J. Heise, Ziad Gemayel, Mohamed Mouine
  • Patent number: 11922326
    Abstract: Approaches are described for generating suggestions for new nodes or new relationships in a knowledge graph based on content of data assets represented by existing nodes in the knowledge graph. The knowledge graph is defined by nodes connected by edges. A method includes determining that a data asset represented by a root node of a knowledge graph has been changed, where the changed data asset is represented by a version node connected to the root node. The changed data asset is processed, including: identifying one or more candidate terms in the changed data asset, and comparing each candidate term with each of one or more existing terms from data assets of the knowledge graph other than the changed data asset to obtain (i) one or more of the candidate terms that do not correspond to any existing term or (ii) one or more candidate terms that each corresponds to a respective existing term that is not related to the version node representing the changed data asset.
    Type: Grant
    Filed: August 26, 2022
    Date of Patent: March 5, 2024
    Assignee: BackOffice Associates, LLC
    Inventors: Kyl Wellman, Jon Green, Tyler Warden, James Maniscalco, Rex Ahlstrom
  • Patent number: 11922332
    Abstract: Embodiments are directed to managing data correlation over a network. Role success models that correspond to roles and to success criteria may be provided. A student profile that includes skill vectors may be provided based on student information. Role success models may be employed to determine intermediate scores based on the skill vectors and the success criteria. A predictive score for the student that corresponds with a predicted performance of the student in the roles may be generated based on the one or more intermediate scores. Actions for the student may be determined based on a mismatch of the skill vectors and role skill vectors that correspond to the roles. In response to the student performing the actions: updating the one or more skill vectors based on a completion of the actions; and updating the predictive score based on the role success models and the updated skill vectors.
    Type: Grant
    Filed: July 26, 2021
    Date of Patent: March 5, 2024
    Assignee: AstrumU, Inc.
    Inventors: Adam Jason Wray, Kaj Orla Peter Pedersen, Xiao Cai, Jue Gong
  • Patent number: 11915122
    Abstract: Methods, systems, and apparatus related to dynamic distribution of an artificial neural network among multiple processing nodes based on real-time monitoring of a processing load on each node. In one approach, a server acts as an intelligent artificial intelligence (AI) gateway. The server receives data regarding a respective operating status for each of monitored processing devices. The monitored processing devices perform processing for an artificial neural network (ANN). The monitored processing devices each perform processing for a portion of the neurons in the ANN. The portions are distributed in response to monitoring the processing load on each processing device (e.g., to better utilize processing power across all of the processing devices).
    Type: Grant
    Filed: July 29, 2020
    Date of Patent: February 27, 2024
    Assignee: Micron Technology, Inc.
    Inventors: Poorna Kale, Amit Gattani
  • Patent number: 11860971
    Abstract: According to an embodiment of the present invention, an approach accurately detects anomalies or outliers of a time-series dataset. A method for identifying whether a particular data element of the time-series dataset is an outlier comprises predicting a value for that particular data element and obtaining a threshold value that defines, relative to the predicted value, whether an actual value of the data element is an outlier. In an aspect of a present invention embodiment, the threshold value is generated based on historic error values associated with data elements temporally preceding the particular data element of the time-series dataset.
    Type: Grant
    Filed: May 24, 2018
    Date of Patent: January 2, 2024
    Assignee: International Business Machines Corporation
    Inventors: Teodora Buda, Hitham Ahmed Assem Aly Salama, Bora Caglayan, Faisal Ghaffar
  • Patent number: 11823015
    Abstract: In various embodiments, a pattern-based recommendation subsystem automatically recommends workflows for software-based tasks. In operation, the pattern-based recommendation subsystem computes an expected distribution of frequencies across command patterns based on different distributions of frequencies across the command patterns. The expected distribution of frequencies is associated with a target user, and each different distribution of frequencies is associated with a different user. The pattern-based recommendation subsystem then applies a set of commands associated with the target user to a trained machine-learning model to determine a target distribution of weights applied to a set of tasks. Subsequently, the pattern-based recommendation subsystem determines a training item based on the expected distribution of frequencies and the target distribution of weights. The pattern-based recommendation subsystem generates a recommendation that specifies the training item.
    Type: Grant
    Filed: January 14, 2019
    Date of Patent: November 21, 2023
    Assignee: AUTODESK, INC.
    Inventors: Tovi Grossman, Benjamin Lafreniere, Xu Wang
  • Patent number: 11797891
    Abstract: The instant systems and methods are directed to a contextual bandits machine learning model configured to enable granular synchronized ecosystem personalization and optimization. The system and methods determine an objective and feed the objective and one more lifecycle model propensity scores as inputs to the contextual bandits machine learning model. The contextual bandits machine learning model then generates one or more potential weighted model rewards, wherein each potential weighted model reward includes at least a desired user action, a weight, a channel, and an expected change to the objective, and selects a weighted model reward that optimizes the objective. An action recommendation is subsequently transmitted to a user device based on the weighted model reward, wherein the action recommendation is presented in a selected channel associated with the weighted model reward. Feedback associated with the action recommendation is collected and used in training and fine-tuning of the model.
    Type: Grant
    Filed: October 24, 2022
    Date of Patent: October 24, 2023
    Assignee: INTUIT INC.
    Inventors: Yashwanth Musiboyina, Dawn-Marie Chantel Miesner, Mustapha Harb, Nan Jiang, Shahram Mohrehkesh, Zachary Dorsch, Suman Sundaresh, Grace Wu
  • Patent number: 11769049
    Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for training a neural network system used to control an agent interacting with an environment to perform a specified task. One of the methods includes causing the agent to perform a task episode in which the agent attempts to perform the specified task; for each of one or more particular time steps in the sequence: generating a modified reward for the particular time step from (i) the actual reward at the time step and (ii) value predictions at one or more time steps that are more than a threshold number of time steps after the particular time step in the sequence; and training, through reinforcement learning, the neural network system using at least the modified rewards for the particular time steps.
    Type: Grant
    Filed: September 28, 2020
    Date of Patent: September 26, 2023
    Assignee: DeepMind Technologies Limited
    Inventors: Gregory Duncan Wayne, Timothy Paul Lillicrap, Chia-Chun Hung, Joshua Simon Abramson
  • Patent number: 11704572
    Abstract: Techniques for selectively offloading data that is computed by a first processing unit during training of an artificial neural network onto memory associated with a second processing unit and transferring the data back to the first processing unit when the data is needed for further processing are described herein. For example, the first processing unit may compute activations for operations associated with forward propagation. During the forward propagation, one or more of the activations may be transferred to a second processing unit for storage. Then, during backpropagation for the artificial neural network, the activations may be transferred back to the first processing unit as needed to compute gradients.
    Type: Grant
    Filed: October 17, 2018
    Date of Patent: July 18, 2023
    Assignee: Zoox, Inc.
    Inventors: Ethan Miller Pronovost, Ethan Petrick Dreyfuss