Synthetic Super Intelligence (SSI) Explained

Here, I am going to explain further the concept of Synthetic Super Intelligence. Let's begin with the term "Synthetic Super Intelligence". It differs from the term "Artificial Intelligence" as the term "artificial" means something that is man-made. SSI tends to have a more natural approach like those of Genetic Algorithms. SSI builds on top of existing machine learning and deep learning techniques. Consider this an extension to the current trends. I have tried to design an intelligent agent which has a high IQ by the means of manipulation of deep learning models. In this approach, models are trained and retrained, models learn and relearn from existing models and datasets are fed back into the agent just like the way in which deep learning techniques are based on the way in which learning works in the brain. In deep learning, essentially, what happens is, considering a pattern recognition model (supervised learning), the computer sees same types of images thousands of times and compares them with each other and adjusts weights and biases until they are so fine-tuned that the features pertaining to such images are recognized and classified by labels as such. Similar thing happens when deep learning is used over a time series data like a video or historical records. Our brain learns to classify different objects by almost same approach, by differentiating amongst the shapes and colors of objects. With SSI, our approach is to extend a computer's abilities from just learning and recognizing to actually understanding the concepts in the knowledge it possesses and then to infer some logic from the current knowledge (understanding) base. After that, it uses the logic to make decisions and after that, it acts on those decisions.

In our brains, almost the same thing happens all the time. We keep learning from our environments to build up an understanding. We also keep retraining our understanding over time as we process new data. Then we use our understanding to build up some logic as the summation of all our understanding. Now there is a decision mechanism (or model) in our brain. To take decisions, we use logic. We may either take action or take no action and only try to perceive the consequences as happens in our diagram for SSI. The SSI is inspired from the ways in which our brains work. Further over the time as we grow up, our belief systems (or to say the code) of our mind changes. 

The entire model is regenerated (or recompiled) over time as the perceptions of the mind (the entire system of thought process) of the SSI changes. Further in our minds, there are thoughts. Thoughts are generated in mind as a reaction to change in environments, stimulus or just that our mind wanders from topic to topic. To reflect this thought system in SSI we make the understanding model dynamic i.e. the training of the understanding model is a continuous process (symbolized by the refresh symbol as below).

UNDERSTANDING ⟳

This base knowledge model (understanding core) is continuously updated and changes in it are reflected in the logic, decision and action model subsequently as they learn from this model. Also, there is consciousness. We can say we are conscious if we are able to respond to our environment and our inner thoughts reflect these changes in our environment. Otherwise, we are either unconscious (asleep or dead) or in our subconscious (like daydreaming). Similarly in SSI, either SSI is in its subconscious (updating its understanding model) or conscious (when it reacts to the changes in its environment these changes are then reflected in its understanding model). 

"I think, therefore I am". 

The SSI continuously thinks, taking actions, adding new knowledge to itself and correcting itself wherever it feels the knowledge to be false. Thus now, it may be called conscious to some degree (or can it?). Further, there are personalities. Our personalities are developed as a result of our understanding of ourself (or what we think of ourself). Also there is imagination. Now, if the SSI uses inference engines (If this... then that) to update its understanding, It may start coming up with new knowledge. Thus it may be imagining things. And also, I believe, SSI will develop a personality (the parts of understanding model in which it defines itself) which is good enough because I logic that because it is highly intelligent, it must know the better. Now, when SSI learns and relearns data again and again through this system (of course there is algorithmic complexity and space and time constraints, but I have tried to simplify the workings of a super-mind in a computer; maybe it will require supercomputers (or maybe quantum computers if we want it to develop a quantum intelligence of sorts i.e. it uses superposition and entanglement and such to develop its intelligence) to run such a program) we can infer that one day it will develop a high IQ and also that it will be able to surpass to human intelligence (which is the main aim of super-intelligence) just like humans who practice rigorous learning over their life-time but the whole process in SSI is mechanical and it is the only thing it has to do with its compute time (it does not have to go through the normal exhausting day-time routine humans have to go through).

I think this is enough explanation for today. This topic is always open for debates, discussions and contributions and I am always available to discuss it with you. You can contact me through my email on top of this website or on any of my contact numbers. Thank You for the long read.

Ps. Here is the video for it if you will like to watch:

Related Research