SELF-IMPROVING INTERACTIONS WITH AN ARTIFICIAL INTELLIGENCE VIRTUAL ASSISTANT

A large language model (LLM) is trained to help users interact with a virtual assistant about products for sale. Users can ask questions or give feedback through a website or app interface. The LLM uses a product knowledge base to generate video responses based on user input. When users provide more feedback after watching the videos, this input helps improve the assistant's responses and the product information. The system learns and updates itself by evaluating user feedback, identifying gaps in information, and gathering data from other sources.

PROACTIVELY-GENERATED PERSONALIZED CONTENT CREATION

A system has been developed to create personalized content for communication messages, especially in marketing. It uses Generative Artificial Intelligence (GAI) to automatically generate various types of content, including text, images, audio, and videos. This system is called Feed Generator (FG) or Content Alchemist. It also includes tracking tools that monitor how well the content performs in real-time. Based on this data, the system can suggest improvements to increase engagement and revenue.

SELECTIVE LEARNING OF INFORMATION FOR THE GENERATION OF PERSONALIZED RESPONSES BY A GENERATIVE RESPONSE ENGINE

A generative response engine can learn important information about a user automatically. It remembers facts, preferences, or contexts from conversations without needing the user to tell it to do so. When the user asks questions or gives prompts, the engine uses this learned information to create personalized responses. Additionally, users have the option to ask the engine to forget specific information if they choose. This technology makes interactions more tailored and user-friendly.

AUGMENTED REALITY EVENT SWITCHING

A system can recognize where a user is looking in one location and find a similar view in another location. When certain conditions are met, it can show images or videos from the second location to the user. These images are taken from the second venue's viewpoint. The visual content is then sent to an augmented reality device that the user is using. This device displays the second venue's visuals as an overlay on what the user is currently seeing.

ARTIFICIAL INTELLIGENCE VIRTUAL ASSISTANT WITH LLM STREAMING

An artificial intelligence virtual assistant uses video streaming to help users with product inquiries. Users can speak to the assistant through a special interface, which captures their audio questions. This audio is processed to create data that is sent to a large language model (LLM) for analysis. The LLM generates responses based on product information and these responses are turned into spoken replies using text-to-speech technology. The virtual assistant then presents these replies in a video format, allowing for a natural conversation with the synthetic human host.

METHOD FOR REAL-TIME GENERATION OF EMPATHY EXPRESSION OF VIRTUAL HUMAN BASED ON MULTIMODAL EMOTION RECOGNITION AND ARTIFICIAL INTELLIGENCE SYSTEM USING THE METHOD

A conversational AI system can recognize emotions in real-time using various methods. It has a model server that provides a machine learning-based conversation model. Users can chat with this model while seeing a virtual human that responds to them. The system captures the user's facial expressions during the conversation to understand their emotions. Based on these emotions, the virtual human adjusts its expressions to show empathy and create a more engaging interaction.

WEARABLE DEVICE PROVIDING IMMERSIVE EXPERIENCE AND METHOD OF CONTROLLING SAME

A wearable device offers an immersive experience by using sensors and cameras to interact with the real world. It can recognize objects around the user and understand their posture. Based on this information, the device calculates how immersed the user is in the experience. While the user interacts with virtual objects displayed on the device, it can adjust how these objects appear based on their level of engagement. This technology aims to enhance the user's experience by making it more interactive and responsive to their actions.

PROMPT AUTO-GENERATION FOR AI ASSISTANT BASED ON SCREEN UNDERSTANDING

Large language models (LLMs) and visual-language models (VLMs) can give helpful answers based on how information is organized. Many users struggle to use these models effectively because they may not know how to input their questions properly. By understanding what is shown on the screen, the system can automatically create prompts to help users interact with AI better. It does this by analyzing a screenshot and matching it with related text descriptions of activities. When it finds a match, it generates helpful suggestions in real-time to assist the user with their current task.

MIXED REALITY INTERACTION WITH EYE-TRACKING TECHNIQUES

Mixed reality (MR) allows users to interact with both digital and real-world elements. By using eye tracking, the system can determine where a user is looking. Once a specific area of interest is identified, users can perform actions like zooming or rotating by using other inputs, such as hand gestures or controllers. This combination of eye tracking and additional gestures makes the interaction more intuitive. Overall, it enhances the way people engage with mixed reality environments.

EMOTIVE TEXT-TO-SPEECH WITH AUTO DETECTION OF EMOTIONS

Emotive text-to-speech technology allows a computer to read text aloud in a way that reflects emotions. It starts by analyzing the text generated by an assistant during a conversation to identify the emotional tone. Based on this emotional analysis, the system creates a special representation of the emotion linked to the text. Then, it uses this emotional information to guide a text-to-speech model in producing speech that sounds like it conveys the intended feelings. This means when the computer speaks, it can express emotions just like a human would.