Overview

Hyper Embodied Intelligence (HEI) serves as the foundational protocol for the creation of In Anima's intelligent agents (referred to as X-embodiments, or Xem). HEI provides the framework and necessary tools for the creation, training, personality development, and social intelligence evolution of Xem agents. This protocol also plays a key role in defining how these agents interface with other users, agents and the digital environment at large.

The core objective of HEI is to establish a deep, seamless connection between the physical and digital realms, enabling these two spaces to converge in terms of information, emotion, and assets.

Xem utilizes large-scale general models to acquire world knowledge, while simultaneously integrating a personalized hybrid large model (pHLM) to emulate the physical subject’s unique characteristics—such as appearance, voice, and behavior. This dual connection allows Xem to engage in continuous cycles of information processing, interaction, and feedback improvement.

Xem features emotional perception and the ability to intelligently adjust communication based on emotional shifts, adapting both the tone of the conversation and vocal modulation accordingly. It also has a comprehensive memory system, which retains crucial facts and emotional changes throughout interactions, building an evolving relationship with users. This relationship influences future conversations, with Xem adapting the conversation style and content based on its emotional understanding of the user.

The pHLM is a hybrid model trained using multimodal data, such as text, voice, images, and video, which allows for iterative development based on feedback from the user. It also allows for adjustments based on changes to the physical subject’s characteristics.

The interactions, collaborations, and connections between Xem agents and other entities within the ecosystem form the backbone of a Hyper Social Network (HSN), creating an unbounded flow of knowledge, emotions, and resources across both physical and digital spaces.