An artificial intelligence-powered movie and series assistant system

Abstract

This invention relates to an artificial intelligence-powered movie and series assistant system that informs and reminds viewers about events that occurred in a specific scene of the movie, documentary, or series they are currently watching, events from previous episodes or seasons of the series, or events that have an impact on or are related to the characters, and enables them to answer any related questions they may have when desired.

Description

[0001] AN ARTIFICIAL INTELLIGENCE-POWERED MOVIE AND SERIES ASSISTANT SYSTEM

[0002] Technical Field

[0003] This invention pertains to an artificial intelligence-powered movie and series assistant system that caters to viewers' needs by providing information and reminders about events that unfold in a particular scene of the movie, documentary, or series they are currently watching. It also helps viewers recall and clarify events from previous episodes or seasons that have a significant impact on the characters or events they may have forgotten or mixed up. The system utilizes artificial intelligence to assist viewers in answering their questions regarding these events.

[0004] Technical Area

[0005] As the complexity of TV shows, movies, and series increases with the proliferation of streaming platforms, it becomes increasingly challenging for viewers to remember previous events, relationships, plotlines, and timelines.

[0006] Some viewers resort to summaries, online databases, forums, or reaching out to their social circles for clarity. However, these methods are either time-consuming or result in a fragmented viewing experience, posing a troublesome issue for consumers of media content.

[0007] To provide an example of the aforementioned problem, consider a scenario where a series viewer forgets the profession of a character or how they met another character within the storyline. They would typically rely on social media, episode summaries, online platforms, or phone calls to friends who have watched the series to obtain the desired information. However, these processes result in both time loss and a potential loss of interest due to the wait, which contradicts the principle of uninterrupted viewing. Various developments in the technical field aim to address this issue. One such development is proposed in the United States patent document US 2021 0012222A1, which suggests a system that presents the most suitable answer from a predefined set of responses to user queries regarding educational content videos. Similarly, the United States patent document US 2020 0380030A1 proposes a system that selects the most relevant video from a predefined video set in response to user questions regarding educational content videos. However, these methods fall short as they have a limited pool of responses that do not address viewers' unique or detailed questions. Both systems require a response pool, and the answer the user receives is limited to the responses available within that pool. The creation of this information pool is costly in terms of labor, requires constant updating, and does not guarantee providing the user with the desired answer.

[0008] Detailed Description of the Invention

[0009] An artificial intelligence-powered movie and series assistant system comprises:

[0010] • A multi-media device, such as a television or a similar device, that allows users to watch content like movies and TV shows and enables them to ask questions about specific events in the scenes they are currently watching or any other related events, and observes the answers to their questions.

[0011] • A scene detection module that determines the start and end time intervals of the scenes in the content to be watched by users.

[0012] • A preprocessing module that applies audio, text, and image processing techniques to the scenes detected by the scene detection module, transforming them into a format that the advanced artificial intelligence module can work on.

[0013] • An advanced artificial intelligence module that extracts the transcriptions of speeches, plays a significant role in generating character, object, time, location, description, classification outputs, or representation vectors of the mentioned outputs using raw audio and visual information of the scenes, and / or receives the question asked by the user about the content they are watching as input and produces the answer to the question as output, using one or more generative artificial intelligence models based on convolutional neural networks or transformers, and stores the artificial intelligence outputs in a database and retrieves them based on vector similarity when needed.

[0014] • An informative module that utilizes an advanced artificial intelligence module to generate a representation vector for a user's question about the content they are watching through the application client, retrieves the closest scenes and information from a database generated by the advanced artificial intelligence module, and produces an original response using the representation vectors, which is then transmitted to the application client.

[0015] • An application client that runs on the multi-media device and enables the user to give a command to learn the desired answer regarding the content they watched through sound or text, allowing them to observe the original answer generated by the informative module.

[0016] For example, when a user cannot recall an event that occurred in a previous season of a TV series or cannot remember the location or time of a movie, the user can type their questions about the subject to the application client. The application client processes these questions and forwards them to the advanced artificial intelligence module. The informative module retrieves the relevant scenes from the scene information database created by the advanced artificial intelligence module and generates a suitable response to the user's question. Subsequently, this response is conveyed to the user through the application client. Thus, viewers can obtain quick and accurate information about the topics they are curious about, resulting in a more enjoyable and seamless viewing experience.

[0017] The advantages gained with the invention are listed below;

[0018] 1.) Real-time solutions for content-related queries: Our innovative system provides assistance in real-time without interrupting the viewing experience. This eliminates the need for viewers to research online sources or forums, reducing interruptions and creating a seamless interactive viewing experience. Generative Al-based responses: The system employs generative Al technologies to generate real-time responses to user inputs. Instead of relying on a predefined answer database, this helps accumulate impressive responses to complex and detailed user questions. Comprehensive content analysis: Our invention conducts multi-input content analysis using inputs from text, audio, and visual sources. This enables a more accurate understanding of the content, resulting in more effective and accurate responses. Wide accessibility: The system can be accessed through a smart TV / mobile application, mobile / web browser application, or browser extension. This expands the system's usability across a wide range of platforms and devices. Cost-effectiveness: By adopting an approach based on generative Al models instead of relying on a predefined answer pool, our system may have lower operating costs. Building and updating the answer pool can be costly and timeconsuming, whereas our generative Al model can dynamically generate answers to the user questions.

Claims

CLAIMS1.) An artificial intelligence-powered movie assistant system that includes:• A multimedia device hardware, such as a television or similar structure, that provides users with the ability to watch content such as movies and TV series, allows them to ask questions about events in scenes they are curious about or any other related events, and observes and provides answers to those questions through its user interface,• A scene detection module that determines the start and end time intervals of scenes in the content to be watched by users,• A preprocessing module that preprocesses the scenes detected by the scene detection module, applying audio, text, and image processing techniques to transform them into a format suitable for the advanced artificial intelligence module to work on,And characterized by• At least one advanced artificial intelligence module equipped with one or more generative Al models based on convolutional neural networks or transformer architectures, which utilize the raw audio and visual information of scenes to generate representation vectors of important characters, objects, time, location, as well as labels, descriptions, classification outputs, or representation vectors of the mentioned outputs, and stores the Al outputs in a database and retrieves them based on vector similarity when needed.• An informative module that utilizes an advanced artificial intelligence module to generate a representation vector for a user's question about the content they are watching through the application client, retrieves the closest scenes and information from a database generated by the advanced artificial intelligence module, and produces an original response using the representation vectors, which is then transmitted to the application client.• An application client that operates within a multimedia device and enables users to obtain the desired information about the content they are watching through voice or text commands, facilitating the presentation and observation of the generated unique response to the user.