MetaSoul INC develops an Emotion Processing Unit (EPU) that enables AI and humanoid robots to generate and respond with real-time emotional states, enhancing their interactions with users. This technology addresses the lack of emotional intelligence in current AI systems, allowing for more authentic and empathetic communication in applications such as virtual caregiving, gaming, and customer support.
Funding
Funding not disclosed
Founders
Product
Problem
Current AI and robotic systems lack the ability to understand, process, and respond to emotions in a way that mirrors human interaction, leading to interactions that feel artificial and lack genuine connection. This deficiency limits the potential for AI to be effectively used in applications requiring empathy and nuanced understanding of human emotional states.
Solution
MetaSoul develops an Emotion Processing Unit (EPU) designed to imbue AI and robots with digital sentience, enabling them to exhibit human-like emotional intelligence. The EPU leverages a specialized Large Behavior Model (LBM) deployed in the cloud to create emotional states and synthetic emotions in intelligent machines. Based on a Psychobiotic Evolutionary Theory, the EPU allows AI to respond to stimuli with a range of primary emotions and sensations, adapting its behavior through reinforcement learning. This technology enables real-time appraisal computation and the creation of an Emotion Profile Graph (EPG), allowing AI to develop unique personalities and respond with appropriate facial expressions and voice tones.
Target Audience
The primary target audience includes developers of AI-driven NPCs for games, humanoid robots, virtual caregivers, and companies seeking to enhance customer experience through emotionally intelligent AI chatbots and virtual assistants.
Features
- Emotion Processing Unit (EPU) that synthesizes emotions in AI and robots
- Large Behavior Model (LBM) deployed in the cloud for creating emotional states
- Psychobiotic Evolutionary Theory-based algorithms for generating twelve primary emotions and sensations
- Real-time appraisal computation using reinforcement learning
- Emotion Profile Graph (EPG) for registering and developing emotional associations
- Integration of Myers-Briggs personality types for tailored emotional communication
- Emo Matrix for real-time emotional control of facial expressions and voice tone
- Available as a System-on-Chip (SoC) for autonomous humanoid robotics with containerized deployment
- Emotion recognition classifiers achieve over 86 percent accuracy (ISEAR) in conversations