Design Advanced Synthetic Environments of Security & Defense for Digital Twins, based on Multimodal Neurocognitive AI (Binomial MILSEV-NC Project).

I share with you, my design reflections and R&D work in Neurocognitive AI, applied to virtual environments for military applications, currently called Military Digital Twins . These environments, we have used them for a long time in the military and security world, and are called Distributed Synthetic Environments (DSEs), they are used for operational training, logistics planning, advanced simulation, engineering, maintenance, development of new doctrines, mission training, war games physical, cybernetic and psychological.

Distributed Synthetic Environments are the core of an AI architecture applied to Digital Twins.

Figure-Example of the architecture of a DSE applied to submarine warfare.

The fundamental objective of this type of system is to test new approaches, models and doctrines in a virtual environment that allows reliably verifying new forms of operation and behavior.

The essential criteria of these environments are based on being able to realistically exhibit the UNPREDICTABILITY and VARIABILITY of behaviors in a complex environment, in order to transfer it to reality progressively or in parallel.

Figure-Model LMT "The Druid" of a psychophysical DSE (fatigue rates, stress levels, decisional response, maneuver precision, sensory coordination, etc.), real-time situational awareness of a combined military operation (air, ground and maritime), in a specific operating scenario.

DSEs have seven essential components:

1-The distributed knowledge base.

2-The modeling of the behaviors and abilities of the actors.

3-The modeling of real world scenarios.

4-The modeling of processes and tasks.

5-Sensors with the real world (humans and environments).

6-Reward mechanisms.

7-Learning mechanism.

The key element of the DSEs is FIDELITY.

Fidelity elements in the DSEs:

1-Sensory fidelity.

2-Physical fidelity.

3-Behavioral fidelity.

4-Time fidelity, latencies and delays.

5-Fidelity of the information.

6-Phenomenological fidelity.

7-Decisional fidelity.

Hybrid AI technologies involved in DSEs

Some of the main artificial intelligence technologies, key for a DSEs, are the following:

1-Blackboard Systems: distribution architecture and data communication between intelligent agents.

2-Expert systems: virtual representation system of the behavior of an expert, in the set of situations where his experience is relevant.

3-Case-based problem solving systems: use the documented experience of successful resolution of "case" problems to solve analogous problems.

4-Evolutionary algorithms: they use population structures and evolve according to rules of selection, recombination, mutation and survival, which are called genetic operations.

5-Belief networks: they use particularizations of the conditional probability theorems to relate phenomena observed in the present (evidence) with known phenomena in the past (previous) and with theories of what is going to happen in the future (models).

6-Fuzzy logic systems: propositions that contain fuzzy values and variables, a fact or a consequence have a fuzzy degree of truth or lie, it serves to express complex situations in a simple way while maintaining fidelity.

7-Neural Networks: Bio-inspired distributed computing model that allows the generation of processing nodes (neurons), connections between neurons to which a weight is assigned based on training or their own feedback, which allows discovering and recognizing patterns of behavior in a complex environment.

Current innovation challenges in DSEs:

Modifiability: dynamic expansion of KB domains, ability to work in multiplatform hardware/software environments.

High-fidelity representation: sensory, factual, characterization of actors and the behavior of said actors.

Adaptive skill mechanisms: maintain believable and robust behaviors under external circumstances and different skill levels of the simulated or real operator.

Data Fidelity: reliable data in a timely manner.

Zero latency between virtual and physical actors.

Identified innovation projects in which I am involved (Binomial MILSEV-NC):

All projects are based on the application of model and proven systems of military DSEs based on Strong AI to market environments.

1-Generation of a hybrid AI algorithm testing environment, for Digital Twins platforms (Microsoft Azure AI, IBM Watson, Google AI, META, etc)

2-Creation and testing of SOAR cognitive architectures, as reference environments for the generation of DSE for DT & Metaverses.

3-Development of CBTS (Composable Behavioral Technologies) and BRGs (Behavior Representation Grammar) tools.

3-Generator and simulator of synthetic hybrid actors multiplatform: SKB Actors (Skills & Knowledge Based Actors)

4-Evolution of the JACK Framework model (BDI environment for the creation of intelligent agents based on JAVA/Python).

5-ZERO-LATENCY architectures.

In future articles I will develop into each of the key elements of Hybrid AI for the creation of distributed synthetic environments, the basis of the architecture of a Digital Twins & Metaverse platforms.

Luis Martín "The Druid"-2024