Ingeniation Reach an AGI Technology Feasibility Milestone – December 2018 18 Dec 2018
The Ingeniation team has passed an important milestone to demonstrate feasibility of our AGI enabling technology. The aim of this feasibility demonstration was to identify, implement and test a fundamental intelligence building block, which we call a Holon. The term ‘Holon’ was coined by the author Arthur Koestler in his book ‘The Ghost in the Machine’ and describes the notion of a whole-part, i.e. something which embraces aspects of wholeness and part-ness simultaneously. There are different interpretations of what a Holon really is.
We have adopted the Holon term for our intelligence building block based on our own interpretation which focusses on its key characteristic: our Holon is a ‘scale changer’; figuratively speaking, integrating many of our Holon building blocks into a system just creates a larger Holon. Most importantly, note that our Holon building block carries fundamental intelligence characteristics, meaning that putting more of such ‘smaller’ intelligences together creates a ‘larger’ intelligence.
In our feasibility studies we have experimentally proven our Holon building block has, among others, the following key characteristics:
Autonomy – Self-deterministic and self-sufficient in its fundamental operation
Perpetuality – Has its own continuous processing cycle
Robustness – Robust against stimuli received from an environment
Reactivity – Reacts spontaneously to changes in stimuli
Adaptability – Capable to adjust its own processing cycle in real-time
Predictiveness – Genuine non-algorithmic predictive capacity
Recognition – Responds to stimuli in a deterministic fashion
Teleologic – Adaptation is driven by an intrinsic goal
To experimentally prove the presence of these and other characteristics in our Holon building block we have developed a comprehensive simulation environment which we call the Holometric Engine. So far, our experimental results indicate that any of these Holon characteristics is, beyond reasonable doubt, a proven fact.
After achieving feasibility, we have started to extend the functionality of the Holometric Engine significantly to conduct experimentation on a larger scale and to target more complex scenarios and specific applications for benchmarking purposes.
First steps in this direction included anomaly detection experiments based on randomly generated binary time series of certain lengths which were used to stimulate a set of Holons in a cyclic fashion. We tested time series length up to 256 bit and without exception Holons were capable to detect anomalies introduced through arbitrary modifications in such time series and adjust to them in real-time. We are confident to say that the Holon is an extremely effective anomaly detector, with the additional capability of spontaneously re-adjusting its processing cycle due to stimuli changes.
Ultimately the Holometric Engine will be iteratively further developed into a full scale commercial AGI enabling platform which Ingeniation will make available to AI solution developers and researchers to support implementation of their own applications. The Ingeniation Technology is a mind-like self-coding cognitive architecture, requiring no programming algorithms to operate.