利索-全球专利检索

  1. Login
  2. Sign Up

Search Results

115 records (took 0.022 seconds)

    USING LANGUAGE MODELS TO IMPROVE THE EXPLAINABILITY OF MACHINE LEARNING SYSTEMS
    1.
    发明申请
    USING LANGUAGE MODELS TO IMPROVE THE EXPLAINABILITY OF MACHINE LEARNING SYSTEMS 有权

    公开(公告)号：US20240394549A1

    公开(公告)日：2024-11-28

    申请号：US18676340

    申请日：2024-05-28

    Applicant: Zeta Global Corp.

    Inventor： Zachary Jones ,  Danny Portman

    IPC: G06N3/0895

    Abstract: The subject technology uses a bootstrapping approach to train language models (LMs) to explain outcomes determined by neural networks, ensemble models, reinforcement learning models, LMs, and other black box machine learning models. The bootstrapping approach may train multiple iterations of a tuned LM using training data determined from progressively complex machine learning models and progressively detailed natural language explanations. The model explanations determined by the tuned LM may be displayed in a user interface (UI) included in a publishing system to provide users more context about and a greater understanding of the decision making process used by the black box machine learning models to determine outcomes.

    NETWORKED COMMUNICATION SYSTEM WITH DATA DEOBFUSCATION LAYER
    2.
    发明公开
    NETWORKED COMMUNICATION SYSTEM WITH DATA DEOBFUSCATION LAYER 审中-公开

    公开(公告)号：US20240144319A1

    公开(公告)日：2024-05-02

    申请号：US18384853

    申请日：2023-10-28

    Applicant: Zeta Global Corp.

    Inventor： Meziane Saidi ,  David Hanzelka ,  Pavan Korada

    IPC: G06Q30/0251 ,  G06Q30/0241

    CPC classification number: G06Q30/0254 ,  G06Q30/0277

    Abstract: The subject technology identifies obfuscated email events received from one or more internet service providers (ISPs). The data deobfuscation layer may identify email messages including obfuscated open events and locations by monitoring the open rates of email messages received by different operating systems, ISPs, and/or device types. The data deobfuscation layer may determine accurate campaign level metrics and/or user open probabilities for batches of email messages having obfuscated events. For example, one or more machine learning models may predict an email open rate for one or more email campaigns and identify the users having the highest probability of generating a true open event. The data deobfuscation layer may be used to improve the performance of email communication networks and/or increase engagement metrics for media campaigns.

    Dynamic content delivery via email
    3.
    发明授权
    Dynamic content delivery via email 有权

    公开(公告)号：US11909701B2

    公开(公告)日：2024-02-20

    申请号：US16460492

    申请日：2019-07-02

    Applicant: Zeta Global Corp.

    Inventor： Christian Monberg ,  Barney Govan ,  Tim Hudson

    IPC: H04L51/18 ,  H04L67/568 ,  H04L51/063

    CPC classification number: H04L51/18 ,  H04L51/063 ,  H04L67/568

    Abstract: Dynamic content can be delivered via email. Email messages include HTML content with one or more hyperlinks referring to a network-connected dynamic content server. The hyperlinks are accessed upon rendering of the email on a client device, triggering a request to a dynamic content server. The dynamic content server determines a content item to be served in response to the request. The dynamic content item determination may be based on factors including the identity of the requester, the email communication in which the hyperlink is embedded, recent activities or behavior on the part of the requester, and/or known preference or demographic information associated with the requester. The request can be directed to a selected content item stored within a content delivery network, for transmission back to the requester and display to the user within the email.

    Personalized content system
    5.
    发明授权
    Personalized content system 有权

    公开(公告)号：US11762927B2

    公开(公告)日：2023-09-19

    申请号：US16460398

    申请日：2019-07-02

    Applicant: Zeta Global Corp.

    Inventor： Christian Monberg ,  Barney Govan

    IPC: G06F16/9535 ,  G06F16/9532 ,  H04L67/53

    CPC classification number: G06F16/9535 ,  G06F16/9532 ,  H04L67/53

    Abstract: Processes and apparatuses for content personalization are provided, providing for rule configuration. Content personalization systems interpret user behavior and attributes along with the content users are interacting with, to build optimized predictive models of what content the user may want to see next. Those predictive models can be utilized to personalize content in one or more environments, including email, mobile and applications. Rules include filters applied to the predictive model output and/or the overall system output. Filters can be prioritized and iteratively applied or removed to adjust system output to satisfy desired result set criteria.

    Audience generation using psychographic data
    6.
    发明授权
    Audience generation using psychographic data 有权

    公开(公告)号：US11682050B2

    公开(公告)日：2023-06-20

    申请号：US17454532

    申请日：2021-11-11

    Applicant: Zeta Global Corp.

    Inventor： Ron Sadi ,  Bashir Eghbali ,  Peter Ridge ,  R. C. Rizza ,  Neej Gore

    IPC: G06Q30/02 ,  G06Q30/0241 ,  G06F40/40 ,  G06Q30/0251 ,  G06Q30/0204

    CPC classification number: G06Q30/0276 ,  G06F40/40 ,  G06Q30/0204 ,  G06Q30/0271

    Abstract: In some examples, a system for determining a level of self-image congruence comprises at least one processor and a memory storing instructions which, when executed by the at least one processor among the processors, cause the system to perform operations comprising, at least: selecting one or more self-image attributes that align with a brand-user image; defining each of the selected self-image attributes via a type of subject matter that is of interest to consumers identifying with the selected attributes; identify consumers exhibiting one or more of the selected attributes based on a digital footprint including an online content consumption; computing a congruence score for each identified consumer relative to each selected attribute; and rating the resulting set of consumers based on their respective congruence scores for the selected attributes.

    DYNAMIC CONTENT DELIVERY
    7.
    发明申请
    DYNAMIC CONTENT DELIVERY 有权

    公开(公告)号：US20220353218A1

    公开(公告)日：2022-11-03

    申请号：US17865022

    申请日：2022-07-14

    Applicant: Zeta Global Corp.

    Inventor： Stefan Beeman ,  Gordon Michael Green ,  Christian Monberg ,  Neeraj Gore

    IPC: H04L51/063 ,  H04L51/10 ,  G06Q30/02 ,  H04L67/02 ,  H04L51/48 ,  H04L51/212 ,  H04L67/568

    Abstract: Dynamic content can be delivered via email. Email messages include HTML content with one or more hyperlinks referring to a network-connected dynamic content server. The hyperlinks are accessed upon rendering of the email on a client device, triggering a request to a dynamic content server. The dynamic content server determines a content item to be served in response to the request. The dynamic content item determination may be based on factors including the identity of the requester, the email communication in which the hyperlink is embedded, recent activities or behavior on the part of the requester, and/or known preference or demographic information associated with the requester. The request can be directed to a selected content item stored within a content delivery network, for transmission back to the requester and display to the user within the email.

    MACHINE LEARNING MODEL AND ENCODER TO PREDICT ONLINE USER JOURNEYS
    9.
    发明申请
    MACHINE LEARNING MODEL AND ENCODER TO PREDICT ONLINE USER JOURNEYS 有权

    公开(公告)号：US20220309117A1

    公开(公告)日：2022-09-29

    申请号：US17704872

    申请日：2022-03-25

    Applicant: Zeta GLobal Corp.

    Inventor： Danny Portman ,  Zachary D. Jones

    IPC: G06F16/9535 ,  G06F11/34 ,  G06N3/04

    Abstract: The subject technology identifies a series of journey event types in an online user journey, the event types including an impression event, an email event, a click event, and a website visit, and assigns an encoder to each event type. Using an assigned encoder, the technology encodes each event type to generate an encoded vector for each event type. The encoded vector is representative of at least a portion of the online user journey relating to that event type. The technology generates an encoded vector for each event type to create a set of encoded vectors, the set of encoded vectors including one or more of an impression event encoded vector, an email event encoded vector, a click event encoded vector, and a website visit encoded vector. The technology aggregates the set of encoded vectors to generate an output of the online user journey encoder, the output including a composite encoded user journey vector for modeling, transmits the output of the online user journey encoder to a user journey training model for training of the model and, using a trained model, generates an occurrence probability for at least one further event in the online user journey.

    ADAPTIVE REAL TIME MODELING AND SCORING
    10.
    发明申请
    ADAPTIVE REAL TIME MODELING AND SCORING 有权

    公开(公告)号：US20220092635A1

    公开(公告)日：2022-03-24

    申请号：US17538647

    申请日：2021-11-30

    Applicant: Zeta Global Corp.

    Inventor： Pavan Korada ,  Sunpreet Singh Khanuja ,  Yun Sam Chong ,  Bharat Goyal ,  Edward Robert Rau, JR.

    IPC: G06Q30/02 ,  G06F16/9535 ,  G06F16/2457 ,  G06F30/20 ,  G06Q30/00

    Abstract: Systems, methods and media for adaptive real time modeling and scoring are provided. In one example, a system for automatically generating predictive scoring models comprises a trigger component to determine, based on a threshold or trigger, such as a detection of new significant relationships, whether a predictive scoring model is ready for a refresh or regeneration. An automated modeling sufficiency checker receives and transforms user-selectable system input data. The user-selectable system input data may comprise at least one of email, display or social media traffic. An adaptive modeling engine operably connected to the trigger component and modeling sufficiency checker is configured to monitor and identify a change in the input data and, based on an identified change in the input data, automatically refresh or regenerate the scoring model for calculating new lead scores. A refreshed or regenerated predictive scoring model is output.

Patent Agency Ranking