technology
The foundational idea of heukno is a self-learning methodology with applications to essentially all technical domains, e.g., quantum computing, optimization, electrical circuit design, material science, drug discovery, and much more. The methodology distinguishes itself in two important ways from large language models (LLMs) and reinforcement learning (RL) techniques:
Instead, the computer explores the domain of interest autonomously. The methodology has the added benefit of allowing explainability of the output. Explainability is a highly desirable quality, but elusive to standard LLMs.
Our name
Our approach replaces the standard neural network heavy approach with one which more closely resembles higher-level human cognitive processes. Small nuggets of insight ('heuristics') are refined and compiled to ever larger structures: heukno = heuristic knowledge.