Why do we find JARVIS, Mr. Spock, Commander Data or C3P0 so lovable? We have all known people like them: they are the ones who see the world in black and white: very left-brained. Spock had an excuse – he was a Vulcan with diminished emotional capacity. Computers, on the other hand, can be downright exasperating when they get stuck in stupid patterns that require us to make everything black or white, exactly right, then spit out garbage when it’s not.
That’s why prompt engines have become an important accessory for Large Language Model (LLM) AI. Human communication, language and cognition is fuzzy by nature, just as much in nature is fuzzy by design. The brain’s capacity for knowledge and the physical mechanisms of thought and understanding are becoming better understood, as are the more abstract realms of hopes and dreams.
The mechanics of electrical transmission in the human neural network support fuzzy processes and outcomes. Yet there seem to be missing links in our understanding of consciousness and judgment. Enter the Mind. Does it exist? What is it made of? How does it interact with the brain? We have few proven answers for these questions but there is a scientific basis for this vexing dichotomy. I will also probe into several other areas of cognition and theory, including the implications of “Threshold Logic” to fuzzy logic.
Empathi Systems use fuzzy logic and threshold functions to process language in much more controlled ways than neural networks like Large Language Models (LLMs). This includes genetic algorithms that mimic the threshold functions of spreading activation in the brain. It also includes forward-chaining and backward-chaining inference capabilities.
Our techniques use an explicit knowledge network as the knowledge base – not a neural network, but a knowledge graph in which each node is a fully formed proposition with a truth or confidence value.
Our process involves specialized areas of the “brain” that resolve contextual components of TIME, SPACE, IDENTITY, CAUSALITY and a dozen other keys of true understanding. That’s why we can confidently provide business users with actionable answers and insights and alidated bias-free explanations.
The difference is in that we do not train an LLM on patterns, we read. Empathi Deep Language Understanding (DLU) understands everything it reads the 1st time, and when it finds information that is not yet known, it adds it to the knowledge network for later validation and curation.
• Our learning adds to its knowledge every hour
• Bibliographic data lineage: you know the source
• Can identify bias, deception and fake news
• Can identify and understand metaphor
• Works with – Does not replace other AI techniques
• Whole-brain approach – Left and Right
