top of page

"Augmented Human Intelligence" vs "Artificial Intelligence"

To explain the principles of "Augmented Human Intelligence", let's make a parallel with Artificial Intelligence. There exist several forms of Artificial Intelligences, and for the sake of argument, let's consider the well-known Neural Network.

neuralnetworks.png

Neural Networks can take several shapes, all composed of an input layer, an output layer and, in the middle, intermediate layers. The learning phase consists in weighting every "connection" to reach the desired output value(s) from a set of input values.

AI can work mostly when it learns from a "big" database. For example, learning how to recognize a dog requires several thousand of dog pictures and non-dog pictures. But this approach has limits: when an AI learns how to differenciate dogs from wolves, it learn that the background is one of the most important parameter. A wolf is still photograph on a white or green background... 

Moreover, when one is trying to engineering something new, by construction, we have NO database. So AI techniques cannot be applied in this context, and this is when Augmented Human Intelligence proves useful.

Augmented Human Intelligence is the elicitation and consolidation of the engineering know-how necessary to build the missing database to learn and thus to fill the missing gap in the value chain.

AHI is the consolidation of know-how from your experts

Geeglee® catches know-how coming from your colleagues, your experts... It's a way to capture the value of your company.

Network.png

We all have patterns in mind, we never do something as total fluke. 

The above figure illustrates an example of know-how consolidated into Geeglee. The blue bubbles represent input layers when the yellow ones represent the way of thinking, up to the right-most one which represents the objective.

Why are patterns of know-how so powerful?

Patterns are common between several Systems-of-Interest.

patterns.png

The "way of thinking" of a martian drone battery and an automotive battery is the same (even if we can size them from a power or an energy perspective). The results differ because we use smaller battery cells for the drone while we use larger battery cells for the car and also because environmental constraints are not the same. Yet, the underlying design logic remains the same.

How to use Geeglee to build such patterns?

Most of our clients build their own patterns: the best way to formalize and consolidate their know-how, make it understandable to everyone in the company and leave it maintainable. Moreover, you can use a standard patterns provided by Geeglee's system engineering team or the use Artificial Intelligence for very complex patterns.

GPatt.png
bottom of page