18364982. LEARNING DIRECTABLE VIRTUAL AGENTS THROUGH CONDITIONAL ADVERSARIAL LATENT MODELS simplified abstract (NVIDIA Corporation)
Contents
LEARNING DIRECTABLE VIRTUAL AGENTS THROUGH CONDITIONAL ADVERSARIAL LATENT MODELS
Organization Name
Inventor(s)
LEARNING DIRECTABLE VIRTUAL AGENTS THROUGH CONDITIONAL ADVERSARIAL LATENT MODELS - A simplified explanation of the abstract
This abstract first appeared for US patent application 18364982 titled 'LEARNING DIRECTABLE VIRTUAL AGENTS THROUGH CONDITIONAL ADVERSARIAL LATENT MODELS
The abstract describes a Conditional Adversarial Latent Model (CALM) process that generates reference motions from original movements to create a library of new movements for an agent, such as virtual characters, animals, or objects. The process involves an encoder mapping requested movements onto a latent space, a low-level policy producing joint movements, a conditional discriminator providing feedback, and a high-level policy controlling macro movements.
- CALM process generates new movements for agents from original reference motions
- Encoder maps requested movements onto a latent space
- Low-level policy produces joint movements for the agent
- Conditional discriminator provides feedback to the low-level policy
- High-level policy controls macro movements for the agent
- High-level policy can utilize rewards or finite-state machine functions
Potential Applications: - Animation and gaming industries for creating diverse movements for characters - Robotics for generating varied motions for robots - Virtual reality for enhancing immersive experiences with lifelike movements
Problems Solved: - Limited movement variety in animations and robotics - Lack of efficient methods to generate diverse motions for agents - Difficulty in controlling and coordinating complex movements in virtual environments
Benefits: - Expanded library of movements for agents - Enhanced realism and diversity in animations and robotics - Improved user experience in virtual reality environments
Commercial Applications: Title: "Enhancing Motion Diversity in Animation, Robotics, and Virtual Reality" This technology can be used in animation studios, robotics companies, and virtual reality developers to create more realistic and diverse movements for their products. It can lead to enhanced user engagement and satisfaction, ultimately driving sales and market growth.
Questions about the technology: 1. How does the CALM process improve the efficiency of generating new movements for agents? 2. What are the key advantages of using a conditional adversarial latent model in motion generation?
Original Abstract Submitted
A conditional adversarial latent model (CALM) process can be used to generate reference motions from a set of original reference movements to create a library of new movements for an agent. The agent can be a virtual representation various types of characters, animals, or objects. The CALM process can receive a set of reference movements and a requested movement. An encoder can be used to map the requested movement onto a latent space. A low-level policy can be employed to produce a series of latent space joint movements for the agent. A conditional discriminator can be used to provide feedback to the low-level policy to produce stationary distributions over the states of the agent. A high-level policy can be employed to provide a macro movement control over the low-level policy movements, such as providing direction in the environment. The high-level policy can utilize a reward or a finite-state machine function.
