Technology

Ingeniation’s breakthrough invention, the Holometric Engine (HE), represents a paradigm shift in artificial intelligence data processing and computation.

Holographic AI – The Holometric Engine (HE)

The HE processes data completely differently to conventional explicitly programmed machine learning and neural network architecture systems. The HE genuinely processes data non-algorithmically, the HE learns autonomously,and no model-training is required.

The totally unique and complex core HE invention displays quantum-like behaviour of superposition and a form of entanglement, and is inspired by theories of harmonic resonance, wave mechanical, fractal intelligence, autopoietic systems, cybernetics, practopoiesis, and information theory.

The HE technology identifies complex patterns in vast volumes of data with minimal pre-programming time, the no specific processing code is more data efficient and accurate, the HE can self-train and process without any domain specific training models.

The Holometric Engine is designed to facilitate; adoption of advanced AI analytics for superior optimization, improved accuracy, and better pattern recognition and prediction, for anomaly and signal detection, clustering, segmentation, video, and image recognition purposes.

The revolutionary HE architecture provides a powerful real time AI processing capability enabling the opportunity to “solve particular complex problems that conventional AI architecture cannot solve”, and facilitating significantly reduced pre-programming and training time costs dependant on the specific use case / application.

Conventional Processing System

Memory created by interacting neurons
Data flows in a series of pre-layered neurons
Construction based on mathematical model
Quantum of weights determine neuron activity
New model design required for separate problems
Model required to be trained with massive dataset
System efficiency is relative to model pre-training

Neural Network Architecture

Holographic Processing System

Instant pattern memory is created by the virtual Processor
Data flows within each holon’s octaplex (24 cell) structure
System construction based on holographic principles
Oscillation of entities (holons) seek harmonic resonance
Quasi-Quantum processing effects faster processing
Model self learns, no external training datasets required
System efficiency dependant on achieving harmonic resonance

Holographic AI Virtual Processor

How Is Holographic AI Different?

Developed Over 19 Years

2002

Revelation of AI Intelligence Theory.

2003

First development attempts.

2007

Discovered the intelligent cell. (like digital units)

2008

Concept Holographic AI Engine designed

2010

The ‘Impulse Cycle’ and Holographic principle introduced.

2014

Ingeniation Pty Ltd established 2014. Holographic AI Engine framework designed.

2017

Perpetual dynamical system developed and tested.

2018

‘Holographic Processing Unit” proof of concept achieved. Patent application submitted for Holographic AI.

2019

Commercial prototype of Holographic AI Engine (HE) developed.

2020

UWA / Chevron / Woodside Subsea Project collaboration. & ICRAR.org Square Kilometre Array Project.

2021

Integrating HE technology into broader use mainstream development platforms (Nvidia Deepstream) to include video and audio.

2002

Revelation of AI Intelligence Theory.

2003

First development attempts.

2007

Discovered the intelligent cell. (like digital units)

2008

Concept Holographic AI Engine designed

2010

The ‘Impulse Cycle’ and Holographic principle introduced.

2014

Ingeniation Pty Ltd established 2014. Holographic AI Engine framework designed.

2017

Perpetual dynamical system developed and tested.

2018

‘Holographic Processing Unit” proof of concept achieved. Patent application submitted for Holographic AI.

2019

Commercial prototype of Holographic AI Engine (HE) developed.

2020

UWA / Chevron / Woodside Subsea Project collaboration. & ICRAR.org Square Kilometre Array Project.

2021

Integrating HE technology into broader use mainstream development platforms (Nvidia Deepstream) to include video and audio.

Conventional AI / Algorithmic

What we call AI today operates within a limited range of pre-defined specific tasks, conventional / narrow AI is not really that intelligent as the specifically coded algorithms and statistical computer programming techniques are only as smart as the human programmers.

Conventional AI based processing has limitations as human AI programmers determine the AI systems neural network architecture processing behaviors. The processing logic is limited to the confines of the programming and the system model training dataset.

Conventional AI systems demonstrate quite impressive achievements but face significant scalability problems that cannot be solved by simply adding more data and more computer power. Conventional AI systems require significant re-programming and resources to upscale.

Conventional AI requires direct involvement of AI experts to assist perform specific narrow AI tasks, it needs massive training datasets which can take hours, days, and weeks to train systems, they can be expensive and inefficient, and accuracy can be constrained.

Holographic AI / Non Algorithmic

Holographic AI can perform a large range of analytical tasks, it creates, adapts and evolves its own processing logic as new data is presented to it. The source code defines the core system processing, no specific set of processing instructions is utilized to process data.

Holographic AI self determines its own programming logic effecting an infinite number of different processing pathways. Holographic AI can be applied to multiple industry analytical domains, applications and use cases, without any programming changes.

Holographic AI features a modular architecture enabling adding entities, holons or linking multiple systems together, facilitating scalability to create new types of applications in previously unknown ways. Programming changes are not required to scale up intelligence complexity.

Holographic AI’s quasi-quantum processing logic self adapts to new input data, without requiring training datasets and new programming to work in any specific new domain. The system immediately recognizes any image previously processed via a unique pattern ID.

Holometric Processing Unit - Solution Architecture

The Holometric Engine (HE) is a virtual machine hosted on a standard computer platform. All the Ingeniation source code resides in the HE Framework of which the HE is the core component, creating a virtual processor.

The virtual processor self-constructs a multi-dimensional highly complex oscillatory neural network structure of intelligent holons (oscillators). The holons continually and dynamically freely interact (with no programmed instruction) in virtual quasi-quantum vibrational states.

The holons instantly modify their internal processing logic to achieve harmonic resonance (synchronisation) within their changing environment in a perpetual cyclic feedback loop, identifying new patterns in input data.