Patents Examined by Barbara M Level
  • Patent number: 12783937
    Abstract: In various examples, a representative subset of data points are queried or selected using integer programming to minimize the Wasserstein distance between the selected data points and the data set from which they were selected. A Generalized Benders Decomposition (GBD) may be used to decompose and iteratively solve the minimization problem, providing a globally optimal solution (an identified subset of data points that match the distribution of their data set) within a threshold tolerance. Data selection may be accelerated by applying one or more constraints while iterating, such as optimality cuts that leverage properties of the Wasserstein distance and/or pruning constraints that reduce the search space of candidate data points. In an active learning implementation, a representative subset of unlabeled data points may be selected using GBD, labeled, and used to train machine learning model(s) over one or more cycles of active learning.
    Type: Grant
    Filed: February 2, 2022
    Date of Patent: November 3, 2026
    Assignee: NVIDIA Corporation
    Inventors: Rafid Reza Mahmood, Sanja Fidler, Marc Law
  • Patent number: 12783924
    Abstract: A quantum method receives by a quantum circuit, electronic information from a first block within a blockchain. The quantum method generates, by the quantum circuit a hash for a second block within the blockchain. The quantum method stores the first block and the second block in a distributed ledger.
    Type: Grant
    Filed: April 3, 2023
    Date of Patent: November 3, 2026
    Assignee: Abu Dhabi University
    Inventors: Hichem El Euch, Mohammed Abdellatif Abdelaal Zidan, Montasir Yousof Abdallah Qasymeh, Ashraf Khalil
  • Patent number: 12783945
    Abstract: Disclosed are a model parameter adjusting method and device, a storage medium and a program product. The method includes: performing classification processing on input data by using a classification model obtained through training based on secure multi-party computing, to obtain classification prediction values of the classification model (S201); performing reduction processing on the classification prediction values (S202); performing normalization processing on the classification prediction values subjected to the reduction processing, to obtain a normalization result of the classification prediction values (S203); and updating a parameter of the classification model according to the normalization result of the classification prediction values (S204).
    Type: Grant
    Filed: March 9, 2023
    Date of Patent: November 3, 2026
    Assignee: BAIDU ONLINE NETWORK TECHNOLOGY (BEIJING) CO., LTD.
    Inventors: Qingyou Yang, Kai He, Jinghua Jiang, Jue Hong
  • Patent number: 12762477
    Abstract: Embodiments generate a machine learning (“ML”) model. Embodiments receive training data, the training data including time dependent data and a plurality of dates corresponding to the time dependent data. Embodiments date split the training data by two or more of the plurality of dates to generate a plurality of date split training data. For each of the plurality of date split training data, embodiments split the date split training data into a training dataset and a corresponding testing dataset using one or more different ratios to generate a plurality of train/test splits. For each of the train/test splits, embodiments determine a difference of distribution between the training dataset and the corresponding testing dataset. Embodiments then select the train/test split with a smallest difference of distribution and train and test the ML model using the selected train/test split.
    Type: Grant
    Filed: August 15, 2023
    Date of Patent: October 13, 2026
    Assignee: Oracle International Corporation
    Inventors: Vikas Agrawal, Karthik Bangalore Mani, Krishnan Ramanathan
  • Patent number: 12762455
    Abstract: A method to correct a fault in the application of a Clifford circuit to a qubit register of a quantum computer comprises: (a) receiving circuit data defining the Clifford circuit; (b) receiving additional data identifying one or more measurements belonging to each of a plurality of faces of a lattice; (c) emitting an outcome code based on the circuit data, the outcome code including a series of outcome checks each corresponding to an anticipated error syndrome for the application of the Clifford circuit to the qubit register; and (d) emitting a topological outcome code based on the circuit data, the additional data, and the outcome code, the topological outcome code including a series of check operators that support quantum-error correction via a topological decoder, thereby enabling fault correction in the application of the Clifford circuit to the qubit register.
    Type: Grant
    Filed: August 14, 2023
    Date of Patent: October 13, 2026
    Assignee: Microsoft Technology Licensing, LLC
    Inventors: Adam Edward Paetznick, Nicolas Guillaume Delfosse, Jeongwan Haah, Michael Edward Beverland
  • Patent number: 12731070
    Abstract: An apparatus of training a machine learning model includes a preprocessing module segmenting a document into each sentence and performing tokenization to generate a token sequence for the document, wherein a document representative token representing the document and representative sentence tokens representing each sentence are included in the token sequence for the document, a first training module training the machine learning model to predict an order of sentences in the document, based on the token sequence for the document, and a second training module training the machine learning model to perform document similarity maximization based on the token sequence for the document.
    Type: Grant
    Filed: October 18, 2022
    Date of Patent: September 8, 2026
    Assignee: SAMSUNG SDS CO., LTD.
    Inventors: Hyunjae Lee, Judong Kim, Hyunjin Choi, Bongkyu Hwang, Jaewoong Yun
  • Patent number: 12718116
    Abstract: A classification model is trained with elements from several data sources, with the elements including sensor data mounted in an industrial plant, and with the labels indicating a semantic type for each of the elements. The classification model is retrained with an adaptive learning algorithm implementing active learning and/or incremental learning, until the classification model is capable of mapping each element of the data sources to one of the semantic types. The method and system provide a semantic mapping for sensor data. The automated or semi-automated creation of the semantic mapping loosens the coupling between a domain expert and data scientist, serves as a bridge and reduces workload, speeding up data modeling and data integration steps. It provides inexperienced users with access to domain expertise. Re-use of data models is facilitated, which simplifies further integration and exchange activities. The adaptive learning algorithm provides an incremental enhancement of the classification model.
    Type: Grant
    Filed: September 1, 2021
    Date of Patent: August 25, 2026
    Assignee: SIEMENS AKTIENGESELLSCHAFT
    Inventors: Nataliia Rümmele, Swathi Shyam Sunder
  • Patent number: 12717875
    Abstract: In an embodiment, a statistical analysis tool is applied on a first set of datapoints related to a first variable associated with a dataset. Based on the application of the statistical analysis tool, statistical information related to the first variable is determined. A set of patterns associated with the first set of datapoints is determined, based on the determined statistical information. Thereafter, a first set of predefined templates associated with the determined set of patterns is determined. Further, a natural language model is applied on the retrieved first set of predefined templates and on the determined statistical information. A first textual explanation of the determined set of patterns is determined, based on the application of the natural language model on the retrieved first set of predefined templates and on the determined statistical information. Further, the determined first textual explanation is rendered on a display device.
    Type: Grant
    Filed: February 23, 2023
    Date of Patent: August 25, 2026
    Assignee: Fujitsu Limited
    Inventors: Mehdi Bahrami, Wei-Peng Chen, Mukul Prasad
  • Patent number: 12718110
    Abstract: Transfer learning for a target domain that partially matches classification of a source domain having a ground truth is enabled. A learning device stores a first set subjected to classification and assigned with a ground truth and a second set having a class that partially matches the first set, and executes, based on a first loss function of first data and a second loss function of second data, processing of updating, by the number of updates, an identifier for identifying the first data and the second data when a feature of the first data or a feature of the second data is input. The learning device calculates a similarity at which the first data is similar to the second data by a data selector that calculates the similarity at the last time, and updates the data selector and the first distribution based on the estimated value.
    Type: Grant
    Filed: July 17, 2023
    Date of Patent: August 25, 2026
    Assignee: HITACHI, LTD.
    Inventors: Koki Takeshita, Naoyuki Terashita
  • Patent number: 12694954
    Abstract: The present disclosure provides a method for generating a small molecule based on a pharmacophore model, a device, and a medium. The method includes: using a fully connected graph to represent a pharmacophore model; using a feature extraction model based on a graph neural network to extract a feature vector of the pharmacophore model from the fully connected graph of the pharmacophore model; performing random sampling on a specified latent variable distribution to obtain a latent variable; and inputting the latent variable and the feature vector of the pharmacophore model to a pre-trained decoder to generate a molecule matching the pharmacophore model, where the pre-trained decoder is a decoder obtained by training a variational autoencoder with training samples.
    Type: Grant
    Filed: December 2, 2022
    Date of Patent: July 28, 2026
    Assignee: Central South University
    Inventors: Min Li, Renyi Zhou, Huimin Zhu
  • Patent number: 12664466
    Abstract: It is provided a method comprising: monitoring if a request to train a machine learning sub-model is received from a meta-training host; generating training data; training the machine learning sub-model by at least a first subset of the training data if the request is received and at least the first subset of the training data is generated; checking if a predefined condition related to the machine learning sub-model is fulfilled; providing the trained machine learning sub-model and at least a second subset of the training data to the meta-training host if the condition is fulfilled.
    Type: Grant
    Filed: August 3, 2020
    Date of Patent: June 23, 2026
    Assignee: NOKIA TECHNOLOGIES OY
    Inventor: Stephen Mwanje
  • Patent number: 12664425
    Abstract: The present invention relates to a real time method for detection of an object that enables to define, by means of a thermal camera, objects that are in the possession of people. The present invention particularly relates to a method that enables the detection of objects that are in the possession of people, through different deep learning methods that are subbranches of artificial intelligence using thermal camera images, wherein the images obtained via thermal cameras are processed real time and input into two different deep learning models.
    Type: Grant
    Filed: August 13, 2021
    Date of Patent: June 23, 2026
    Assignee: BAHCESEHIR UNIVERSITESI
    Inventors: Ozan Veranyurt, Cemal Okan Sakar
  • Patent number: 12664473
    Abstract: A computer-implemented method includes generating a candidate machine learning model. The candidate machine learning model is configured to generate probability scores for members. The method includes training the candidate machine learning model using an initial data set, testing the candidate machine learning model to generate performance metrics indicative of a bias value correlated to the candidate machine learning model, and determining whether the performance metrics are below a threshold. In response to determining that the performance metrics are below the threshold the method includes generating a proxy feature based on the bias value, adding the proxy feature to the candidate machine learning model, calculating weights based on the bias value, updating the initial data set using the weights to generate an updated data set, and retraining the candidate machine learning model using the updated data set.
    Type: Grant
    Filed: February 21, 2023
    Date of Patent: June 23, 2026
    Assignee: Evernorth Strategic Development, Inc.
    Inventors: Sandip Sonawane, Rahul Rajendran, Vijay Sai Kondamadugu, Amit K. Bothra
  • Patent number: 12664472
    Abstract: Provided herein are systems and methods for detecting and sanitizing sensitive data using domain-enhanced attention neural networks. In some embodiments, a processor of a client retrieves training data comprising tuples. Each tuple comprises a first parameter and a second parameter. For each tuple the processor matches a substring in a first parameter of the respective tuple to a keyword of a plurality of keywords, identifies a security category corresponding to the at least one keyword, and expands the first parameter of the respective tuple to comprise a respective string associated with the application event and the security category. The processor trains a model to detect and sanitize the sensitive data from the application events using the tuples, including an expanded first parameter for each tuple. The processor sanitizes the sensitive data using the trained model.
    Type: Grant
    Filed: January 9, 2023
    Date of Patent: June 23, 2026
    Assignee: Capital One Services, LLC
    Inventors: Rui Zhang, Sudheendra Kumar Kaanugovi, Sandeep K. Gadde
  • Patent number: 12657515
    Abstract: A non-transitory computer-readable recording medium stores a model generation program for causing a computer to execute a process including: acquiring results output from a first machine learning model in response to input of a first plurality of pieces of data to the first machine learning model; selecting a second plurality of pieces of data from the first plurality of pieces of data, based on the results; and generating a second machine learning model by executing machine learning of the first machine learning model, by using the second plurality of pieces of data as input, with some parameters being fixed among a plurality of parameters included in the first machine learning model.
    Type: Grant
    Filed: February 3, 2023
    Date of Patent: June 16, 2026
    Assignee: Fujitsu Limited
    Inventor: Yoshihiro Okawa
  • Patent number: 12645994
    Abstract: Disclosed is an approach for performing auto-classification of documents. A machine learning framework is provided to analyze the document, where labels associated with certain documents can be propagated to other documents.
    Type: Grant
    Filed: December 28, 2022
    Date of Patent: June 2, 2026
    Assignee: Box, Inc.
    Inventors: Divya Jain, Adelbert Chang, Lance Co Ting Keh, Shivani Rao, Sivaramakrishnan Subramanian
  • Patent number: 12646004
    Abstract: Feature engineering, for example, in automated machine learning, can include receiving streaming data representing at least one attribute detected by a sensor over time. Long term point statistics associated with the streaming data can be computed. The streaming data can be quantized into intervals of time windows and short term point statistics based on the intervals can be computed. The long term point statistics and the short term point statistics can be normalized. Dynamic time warping can be applied across the normalized long term point statistics and short term point statistics. A pair of probability distributions can be generated associated with the dynamic time warped normalized long term point statistics and short term point statistics. Based on distance between the mean values of the probability distributions, machine learning input features can be produced. The machine learning input features can be fed to train a machine learning model for detecting anomaly.
    Type: Grant
    Filed: March 23, 2023
    Date of Patent: June 2, 2026
    Assignee: International Business Machines Corporation
    Inventors: Kunal Sawarkar, Shivam Raj Solanki, Christopher Chen, Amit P. Joglekar
  • Patent number: 12645996
    Abstract: A machine learning model is accessed that is configured to use one or more parameters to process images to generate labels. The machine learning model is executed to transform at least part of each of at least one digital pathology image into a plurality of predicted labels; and generate a confidence metric for each of the plurality of predicted labels. An interface is availed that depicts the at least part of the at least one digital pathology image and that differentially represents predicted labels based on corresponding confidence metrics. In response to availing of the interface, label input is received that confirms, rejects, or replaces at least one of the plurality of predicted labels. The one or more parameters of the machine learning model are updated based on the label input.
    Type: Grant
    Filed: January 31, 2023
    Date of Patent: June 2, 2026
    Assignee: VENTANA MEDICAL SYSTEMS, INC.
    Inventors: Hadley Fellows, Mehrnoush Khojasteh, Justine Larsen, Jim F. Martin, Nidhin Murari, Fahime Sheikhzadeh
  • Patent number: 12632743
    Abstract: In an example embodiment, a random forest machine learning algorithm is used to create and/or identify rules to apply to an individual entity in a computer system that has a plurality of entities, each with a number of rules. More precisely, rule predicates are used as features of a random forest model built to predict a particular outcome (e.g., a transaction that is fraudulent). Hyperparameters of the random forest model are varied and iterated. A classifier is used to calculate feature importance for all features in the training data. Feature importance may be calculated using permutation feature importance. The N “most important” features are then found from this set. The N “most important” features are then used to find rules above a certain precision and recall rate. These rules may then be backtested and the best rules can be used to generate additional rules.
    Type: Grant
    Filed: August 24, 2022
    Date of Patent: May 19, 2026
    Assignee: Stripe, LLC
    Inventors: Ariel David Sagalovsky, Chiranth Manjunath Hegde
  • Patent number: 12626188
    Abstract: An artificial intelligence (AI) feedback method includes: providing input1 and input2 to a script extractor and executor, providing, by the script extractor and executor, an initial condition to an AI, generating, by the AI, a two-dimensional (2D) information table from the initial condition, adding the popularity frequency value or ranking value to corresponding coordinates in the 2D information table and matching knowledge by an interaction between the AI and the script extractor and executor, generating, by the AI, a 2D pseudolinear transformation table from the 2D information table, and performing, by the AI, pseudolinear transformation multiple times to form a 2D unique characteristic table and deriving information of coordinates whose characteristic vector is not changed as a deep truth value.
    Type: Grant
    Filed: August 3, 2021
    Date of Patent: May 12, 2026
    Assignee: KOREA HYDRO & NUCLEAR POWER CO., LTD.
    Inventor: Seung Chan Lee