When we think about Artificial Intelligence (AI) we might think that means:
· machines that think like humans,
· machines that act like humans,
· machines that think rationally,
· machines that act rationally.
The reality is very different, even some of the most developed multi-billion dollar systems, are not able to do the things that the human / animal brain does with little or no difficulty. The problem is that actions such as sight or recognition are difficult to program. In reality, the brain creates numerous constructs to work together to complete a more complex task. The more you do something the more constructs get created and the most efficient group survives.
You might think of this as muscle memory; you do something so many times it becomes natural. You execute without thought and requiring very little energy.
Computer systems do not work like this. The simplicity and elegance of the synapsis is not replicated in today’s technology. The amount of compute power and energy required to run an operation is the same no matter how many times it is executed. In other words, “no muscle memory, or learning”. artificial intelligence machine learning
What we are able to do today is create algorithms. Strung together, they can manipulate multiple data sets or even different types of information and glean some predefined information. Better algorithms can even seem intuitive as the look for outliers of information that would not seem to not fit the data set and identify those for further elaboration.
Of course, more compute power, better algorithms and good data can produce very powerful insights into our client’s thought, our processes, and even our own attitudes. machine learning
Finally, one the greatest challenges, is common sense. Those “learned” norms of what is right and wrong, what is most likely to work, what just never makes sense, and what is acceptable in today’s society. Granted, many humans are challenged when it comes to common sense, but for the computational array, the capability of building relative constructs and “learning” is still a long way off. deep learning
Bringing together experts in systems, networks, communications and programming we can help identify what can be accomplished, multiple hardware and software solutions, vendors, make vs. buy analysis, and how to proceed implementing those plans.