NEAR-MEMORY PSEUDORANDOM NUMBER GENERATION
    1.
    发明公开

    公开(公告)号:US20240311084A1

    公开(公告)日:2024-09-19

    申请号:US18606809

    申请日:2024-03-15

    IPC分类号: G06F7/58

    CPC分类号: G06F7/582

    摘要: Various examples are directed to systems and methods for generating a set of pseudorandom numbers in a computing system comprising a compute element and a memory device. A memory controller of the memory device may receive, from the compute element, an indication to generate a set of pseudorandom numbers. The memory controller may generate the set of pseudorandom numbers and write the set of pseudorandom numbers to a memory array of the memory device for access by the compute element.

    CONTENT INITIALIZATION METHOD, ELECTRONIC DEVICE AND STORAGE MEDIUM

    公开(公告)号:US20240275848A1

    公开(公告)日:2024-08-15

    申请号:US18020618

    申请日:2022-08-01

    IPC分类号: H04L67/1097 G06F7/58

    摘要: The present disclosure provides a content initialization method and apparatus, an electronic device and a storage medium, which relates to a field of computer technology, in particular to fields of deep learning and distributed computing. The content initialization method is applied to any one of a plurality of devices included in a distributed system. A specific implementation scheme of the content initialization method is: determining, according to a size information of a resource space for the distributed system and an identification information of the any one of the plurality of devices, a space information of a first sub-space for the any one of the plurality of devices in the resource space, wherein the space information includes a position information of the first sub-space for the resource space; and determining an initialization content for the first sub-space according to a random seed and the position information.

    HOLD AND SPIN FEATURE GAME WITH MYSTERY SYMBOLS

    公开(公告)号:US20240233490A1

    公开(公告)日:2024-07-11

    申请号:US18612466

    申请日:2024-03-21

    发明人: Daniel Marks Hua Xu

    摘要: A gaming machine provides a base game from which a feature game is triggered when a determined number of configurable symbols are displayed in a base game outcome. When the feature game is triggered, the configurable symbols are held in place on the display and the player is provided with more spins during the feature game in which to collect additional configurable symbols. Any additional configurable symbols are retained on the display during subsequent spins until the feature game is completed. The configurable symbols include mystery symbols whose value is not determined until the mystery symbol is displayed. A first valuation game, such as a wheel spin, is used to determine the value for the mystery symbols. At the end of the feature game, the player is awarded a prize based on the values carried by the configurable symbols. A second valuation game, such as a wheel spin, is used to provide an enhanced award if the player collects a determined number of configurable symbols during the feature game.

    TRAINING AND TESTING UTTERANCE-BASED FRAMEWORKS

    公开(公告)号:US20240203401A1

    公开(公告)日:2024-06-20

    申请号:US18530702

    申请日:2023-12-06

    申请人: Spotify AB

    发明人: Daniel Bromand

    摘要: Systems, methods, and devices for training and testing utterance based frameworks are disclosed. The training and testing can be conducting using synthetic utterance samples in addition to natural utterance samples. The synthetic utterance samples can be generated based on a vector space representation of natural utterances. In one method, a synthetic weight vector associated with a vector space is generated. An average representation of the vector space is added to the synthetic weight vector to form a synthetic feature vector. The synthetic feature vector is used to generate a synthetic voice sample. The synthetic voice sample is provided to the utterance-based framework as at least one of a testing or training sample.

    SYSTEMS AND METHODS FOR RANDOM NUMBER GENERATION

    公开(公告)号:US20240168720A1

    公开(公告)日:2024-05-23

    申请号:US18113735

    申请日:2023-02-24

    发明人: Mohammad H. Amin

    IPC分类号: G06F7/58 G06N10/40 H04L9/08

    摘要: Systems and methods for random number generation are discussed. A first processor is in communication with a quantum processor, the quantum processor having an array of superconducting qubits. The first processor instructs the quantum processor to selectively communicatively couple the superconducting qubits to embed a quantum system having a highly entangled nontrivial ground state. The highly entangled nontrivial ground state comprising a uniform distribution of classical ground states. One or more distortions are introduced to the uniform distribution by one or more random variations based on an input value. The quantum processor evolves over the embedded quantum system. A set of one or more random numbers is received from the quantum processor.

    Method And System For Log Data Analytics Based On SuperMinHash Signatures

    公开(公告)号:US20240039699A1

    公开(公告)日:2024-02-01

    申请号:US18376882

    申请日:2023-10-05

    申请人: Dynatrace LLC

    摘要: A system and method for the analysis of log data is presented. The system uses SuperMinHash based locality sensitive hash signatures to describe the similarity between log lines. Signatures are created for incoming log lines and stored in signature indexes. Later similarity queries use those indexes to improve the query performance. The SuperMinHash algorithm uses a two staged approach to determine signature values, one stage uses a first random number to calculate the index of the signature value that is to update. The two staged approach improves the accuracy of the produced similarity estimation data for small sized signatures. The two staged approach may further be used to produce random numbers that are related, e.g. each created random number may be larger than its predecessors. This relation is used to optimize the algorithm by determining and terminating when further created random numbers have no influence on the created signature.

    Communication Efficient Federated Learning
    10.
    发明公开

    公开(公告)号:US20230376856A1

    公开(公告)日:2023-11-23

    申请号:US18365734

    申请日:2023-08-04

    申请人: Google LLC

    摘要: The present disclosure provides efficient communication techniques for transmission of model updates within a machine learning framework, such as, for example, a federated learning framework in which a high-quality centralized model is trained on training data distributed overt a large number of clients each with unreliable network connections and low computational power. In an example federated learning setting, in each of a plurality of rounds, each client independently updates the model based on its local data and communicates the updated model back to the server, where all the client-side updates are used to update a global model. The present disclosure provides systems and methods that reduce communication costs. In particular, the present disclosure provides at least: structured update approaches in which the model update is restricted to be small and sketched update approaches in which the model update is compressed before sending to the server.