中科微点-全球专利检索

  1. Login
  2. Sign Up

Search Results

1309 records (took 0.081 seconds)

    RULE-BASED TECHNIQUES FOR EXTRACTION OF QUESTION AND ANSWER PAIRS FROM DATA
    3.
    发明申请
    RULE-BASED TECHNIQUES FOR EXTRACTION OF QUESTION AND ANSWER PAIRS FROM DATA 审中-公开

    公开(公告)号：WO2023076756A1

    公开(公告)日：2023-05-04

    申请号：PCT/US2022/074978

    申请日：2022-08-15

    Applicant: ORACLE INTERNATIONAL CORPORATION

    Inventor： ZHONG, Xu

    IPC: G06F16/9032 ,  G06F16/332 ,  G06F16/901 ,  G06F16/955 ,  G06F40/279 ,  G06F40/143

    Abstract: Techniques are disclosed for rules-based techniques for extraction of question-and-answer pairs from digital documents. In an exemplary technique, a digital text document can be accessed by executing a document indicator. A document hierarchy can be generated. The document hierarchy can include at least one parent node and child nodes corresponding to the parent node. Each node of the child nodes can correspond to a text feature of the digital text document. At least a first child node of the child nodes can be determined that corresponds to a first text feature. The parent node of the first child node and a second child node of the parent node can be determined. The second child node can correspond to a second text feature related to the first text feature. A training data set can be generated.

    FINE-TUNING MULTI-HEAD NETWORK FROM A SINGLE TRANSFORMER LAYER OF PRE-TRAINED LANGUAGE MODEL
    4.
    发明申请
    FINE-TUNING MULTI-HEAD NETWORK FROM A SINGLE TRANSFORMER LAYER OF PRE-TRAINED LANGUAGE MODEL 审中-公开

    公开(公告)号：WO2023064033A1

    公开(公告)日：2023-04-20

    申请号：PCT/US2022/040530

    申请日：2022-08-17

    Applicant: ORACLE INTERNATIONAL CORPORATION

    Inventor： VU, Thanh Tien ,  PHAM, Thomas ,  NEZAMI, Omid Mohamad ,  JOHNSON, Mark Edward ,  DUONG, Thanh Long ,  HOANG, Cong Duy Vu

    IPC: G06N3/08 ,  G06N20/00

    Abstract: Techniques are provided for customizing or fine-tuning a pre-trained version of a machine-learning model that includes multiple layers and is configured to process audio or textual language input. Each of the multiple layers is configured with a plurality of layer-specific pre-trained parameter values corresponding to a plurality of parameters, and each of the multiple layers is configured to implement multi‑head attention. An incomplete subset of the multiple layers is identified for which corresponding layer-specific pre‑trained parameter values are to be fine-tuned using a client data set. The machine-learning model is fine-tuned using the client data set to generate an updated version of the machine‑learning model, where the layer‑specific pre-trained parameter values configured for each layer of one of more of the multiple layers not included in the incomplete subset are frozen during the fine-tuning. Use of the updated version of the machine-learning model is facilitated.

    MACHINE LEARNING BASED CRYPTANALYSIS
    5.
    发明申请
    MACHINE LEARNING BASED CRYPTANALYSIS 审中-公开

    公开(公告)号：WO2023049737A1

    公开(公告)日：2023-03-30

    申请号：PCT/US2022/076768

    申请日：2022-09-21

    Applicant: ORACLE INTERNATIONAL CORPORATION

    Inventor： GANESAN, Dharmalingam ,  CLIFTON, David M.

    IPC: G06N20/00 ,  H04L9/00

    Abstract: Embodiments decrypt or partially decrypt an encoded message or a private key, the encoded message or private key encoded by a public-key cryptography algorithm. Embodiments encode the public-key cryptography algorithm using a language of a program synthesizer and construct a grammar for the program synthesizer. Embodiments train the program synthesizer with training data comprising input-output pairs and execute the trained program synthesizer to generate a mathematical formula. Embodiments validate the generated mathematical formula and then perform the decrypting using the trained and validated program synthesizer.

    PREDICTION OF BUFFER POOL SIZE FOR TRANSACTION PROCESSING WORKLOADS
    7.
    发明申请
    PREDICTION OF BUFFER POOL SIZE FOR TRANSACTION PROCESSING WORKLOADS 审中-公开

    公开(公告)号：WO2023003622A1

    公开(公告)日：2023-01-26

    申请号：PCT/US2022/030119

    申请日：2022-05-19

    Applicant: ORACLE INTERNATIONAL CORPORATION

    Inventor： FAIZIAN, Peyman ,  BENCY, Mayur ,  KOCBERBER, Onur ,  SUNDARA, Seema ,  AGARWAL, Nipun

    IPC: G06F16/21 ,  G06F11/34

    Abstract: Techniques are described herein for prediction of a buffer pool size (BPS). Before performing BPS prediction, gathered data are used to determine whether a target workload is in a steady state. Historical utilization data gathered while the workload is in a steady state are used to predict object-specific BPS components for database objects, accessed by the target workload, that are identified for BPS analysis based on shares of the total disk I/O requests, for the workload, that are attributed to the respective objects. Preference of analysis is given to objects that are associated with larger shares of disk I/O activity. An object-specific BPS component is determined based on a coverage function that returns a percentage of the database object size (on disk) that should be available in the buffer pool for that database object. The percentage is determined using either a heuristic-based or a machine learning-based approach.

    ROUTING POLICIES FOR GRAPHICAL PROCESSING UNITS
    8.
    发明申请
    ROUTING POLICIES FOR GRAPHICAL PROCESSING UNITS 审中-公开

    公开(公告)号：WO2022271991A1

    公开(公告)日：2022-12-29

    申请号：PCT/US2022/034789

    申请日：2022-06-23

    Applicant: ORACLE INTERNATIONAL CORPORATION

    Inventor： BRAR, Jagwinder Singh ,  BECKER, David Dale ,  SULEK, Lukasz ,  UECKER, Jacob Robert

    IPC: H04L45/24 ,  H04L45/7453 ,  H04L45/00 ,  H04L45/28 ,  H04L45/38 ,  H04L47/125 ,  H04L49/1569

    Abstract: Discussed herein is a routing mechanism for graphical processing units (GPUs) that are hosted on several host machines in a cloud environment. For a packet transmitted by a GPU of a host machine and received by a network device, the network device determines an incoming port-link of the network device on which the packet was received. The network device obtains a flow information associated with the packet, and computes, based on the flow information, an outgoing port-link of the network device in accordance with a hashing algorithm. The hashing algorithm is configured to hash packets received on a particular incoming port-link of the network device to be transmitted on a same outgoing port-link of the network device. The network device forwards the packet on the outgoing port-link of the network device.

    TECHNIQUES FOR SAFE DATABASE MIGRATION WITH NO DOWNTIME
    10.
    发明申请
    TECHNIQUES FOR SAFE DATABASE MIGRATION WITH NO DOWNTIME 审中-公开

    公开(公告)号：WO2022271437A1

    公开(公告)日：2022-12-29

    申请号：PCT/US2022/032367

    申请日：2022-06-06

    Applicant: ORACLE INTERNATIONAL CORPORATION

    Inventor： LUONG, Duc ,  KOU, Xuekun ,  LONCARIC, Calvin

    IPC: G06F16/21 ,  G06F16/27 ,  G06F16/214 ,  G06F16/2358 ,  G06F16/2379 ,  G06F16/278

    Abstract: Techniques for enabling efficient data migration between data stores with no downtime are disclosed. A distributed computing system can be implemented with an initial data store and a target data store. During the migration of a portion of the data from the initial data store to the target data store, the distributed computing system can receive requests to create data entities or launch workflow instances at the data stores. The system can determine whether the initial data store or the target data store has been designated the primary data store for handling the requests. The system can also determine whether the initial data store or the target data store contain a key associated with the request. If the key is present in either of the data stores, the system may abort the creation of the data entity.

Patent Agency Ranking