Automated Virtual Instructors
Patterns of Life
Creating Realistic Experiences
Artificial Intelligence Unlike Any Other
Discovery Machine, Inc. is an Artificial Intelligence (AI) company unlike any other.
DMI leverages a wide range of AI techniques from knowledge acquisition (KA) to machine learning (ML) to develop “intelligent constructs” for training, decision support and automation. DMI’s highly acclaimed, patented knowledge capture methodology works in conjunction with our patented visual modeling tools to enable the agile production of intelligent constructs. DMI’s AI overcomes the limitations of ML imposed by sparse data environments by capturing the mental models trapped in the heads of your organization’s subject matter experts (SME) to bias and direct learning.
This will be the overview video
Intelligent construct is a term encompassing both intelligent agents and intelligent devices. DMI leverages intelligent constructs to create experiences. These experiences can be used for training or to anticipate operational situations. The RESITE® suite of software enables non-programmers to author experiences by adding, combining and linking constructs into scenarios. DMI intelligent constructs leverage subject matter expertise captured from your organizations’ leading experts, which is then deployed into any scene or setting. DMI intelligent constructs are also independent of any specific simulation or operational environment. DMI uses the Multi-level Universal Specification for Intelligent Constructs (MUSIC) to integrate with simulations or operational environments.
RESITE® Suite of Software
Enables non-programmers to author experiences by adding, combining and linking constructs into scenarios.
Discovery Machine Intelligent Agents are:
- Situationally aware
DMI agents are currently deployed in the following roles:
automated virtual instructors, patterns of life (PoL) agents, adversary agents, teammate agents, and decision support agents.
These agents are also paired with intelligent devices which represent complex systems found in the environment. For example, a virtual instructor can have a mental model of a device which exists in the world such as a natural gas well-head. The intelligent device can be parameterized to exhibit a wide variety of behaviors resulting in, for example, a variety of pressure gauge readings. Insofar as the agent’s mental model captures how the device works, it can then use, diagnose or predict the behaviors of the device. The virtual instructor can also observe trainees to see how they deal with the device and offer assistance when their diagnoses or predictions are incorrect.