Invention Title:

PERSONALIZED REAL-TIME ADVERTISEMENT CONTENT GENERATION

Publication number:

US20250254401

Publication date:
Section:

Electricity

Class:

H04N21/812

Inventors:

Assignee:

Applicant:

Drawings (4 of 7)

Smart overview of the Invention

The patent application describes a system for generating personalized real-time advertisements using advanced technologies. The system predicts a user's most preferred product from a selection of available products by utilizing a trained neural network. This prediction considers user data and the real-time availability of products. The system then generates a script using a large language model, which creates a customized offer for the user. The generated script is used to produce multimedia content tailored to the user, utilizing a generative adversarial network (GAN) to create audio and video elements.

Background

Traditional advertising methods often rely on static content that does not adapt to individual user preferences or real-time data. Dynamic advertisements, however, adjust their content and presentation based on various factors such as user behavior, location, and demographics. This approach enhances user engagement by providing relevant and personalized ad experiences. Dynamic Creative Optimization (DCO) further refines this process by using data-driven insights to tailor ad content in real-time, transforming static campaigns into dynamic interactions that resonate with individual users.

Technical Summary

The system leverages multiple cutting-edge technologies to enhance advertisement personalization. A neural network predicts the user's preferred product by analyzing real-time data and user behavior. A large language model generates a customized offer script, which is then transformed into multimedia content through a GAN. This content is presented on the user's device, ensuring that the advertisement is both timely and relevant. The system's architecture includes components such as processors, memory, and storage media to facilitate these operations.

Detailed Description

Existing DCO systems often fall short in creating complex narratives that adapt to multiple user parameters due to limited data usage. The proposed solution addresses these limitations by integrating diverse data sources like real-time product inventory, user location, and historical data to craft personalized ad narratives. This approach ensures that advertisements are not only personalized but also contextually relevant, adapting dynamically to changing conditions and user preferences.

Implementation Details

The system processes product and user data to generate personalized advertisements. Product data includes specifications, purchase history, and availability, while user data encompasses location, online behavior, demographics, and preferences. Contextual information such as time of day and local events further personalizes the ads. A trained deep neural network processes this information, predicting the user's most preferred product by analyzing temporal patterns in behavior and interactions. The resulting advertisement is tailored in real-time to maximize engagement and relevance.