For a really long time, humans have looked into nature to find answers to challenges out of curiosity and the desire to learn and discover. Scientists, Mathematicians, Physiciatians, physiologists, and Philosophers in the quest to develop new methods and approaches and in the search for classifying, and learning from nature and its complexity, to fuse new knowledge, techniques, and capabilities.
In early 19th century some of the original thinkers of Artificial intelligence, Mathematicians, Psychologies and Physics principles and scientific theories, established and contributed the development of the term Artificial Intelligence, there were, John McCarthy, Alan Turning, Marvin Minsky, Allen Newell, Herbert Simon and Seymour Papert.
In the early 1950’s where John McCarthy defined Artificial Intelligence and its associated programs with the capabilities to learn and reason described in symbolic and mathematical form, following the physics and philosophical theories of Albert Einstein, David Bohn, Martin Minsky, Herbert Simon underpinned by the mathematical and scientific work of Alan Turing and Seymon Peppers to understand and represent the nature and human behaviour around us.
Artificial Intelligence has evolved from it’s long history of AI ranging from searching knowledge frames, analysing human languages, propagating constraints, running neural networks, deep learning and back again to extracting and making vast amount of knowledge useful for normal people in the modern, distributed world, from software robots, industrial robots, and to intelligent design systems for products and services.
In the 21st century where we are moving towards generative and institute systems. AI as a Service makes it much simpler and efficient approach for organisations to develop capabilities, services and products without the need to onboard a big development team nor a deep understanding of AI.
AI as a service is not just learning and classification, also known as the dark side of AI, comprising mostly of mathematical modelling (statistics) also known by their popular names neural networks, deep learning, and machine learning (ML).
Symbolic Languages in AI as a Service
AI as a Service carries forward all of the 6 decades of AI, which includes reasoning and explaining all with symbolic languages – symbolic AI. This is where programming languages good at symbolic data processing shines, Lisp being the popular choice from America, and Prolog the popular choice from Europe and Asia. They are both in active use, and EsseSystems has chosen Lisp as the way forward, because it the most flexible in the various styles necessary for both learning capabilities and reasoning capabilities.
Genetic Algorithms and Ontologies in AI as a Service
Therefore, a field of study was created, from which many technologies used today came from. In this field, called Bio-inspired computing, researchers have observed some aspect of nature, and imitated that, in the form of algorithms and structures, to present answers to problems that maybe couldn’t be solved by conventional means. One of the big examples of Bio-inspired computing are the Genetic Algorithms (GA).
GA’s are based on the theory of evolution. These algorithms try to evolve a population, seeking to present a good answer to some of the challenges faced in our society. For it, they replicate methods of natural selection, crossover, mutation, and more. Consequently, they are vastly used in tasks that involve optimization. And this optimization could be (and have been) used by companies.
It is curious to note that in AiaaS, not much is said about GA’s, the spotlight mostly turned to Machine Learning and Deep Learning algorithms, but it’s safe to assume that they (the GA’s) are present in many applications, as the urge to find good and reliable solutions, in a short amount of time have dominated the market. Having said that, to prosper in a scenario where AiaaS is becoming more common, it is essential to know the available tools, to take your process to the next level, and the Genetic Algorithms are, without a doubt, one of them.
And couldn’t finish this topic without mention the form of how knowledge is stored and organized. In companies, as in many other areas, it is fundamental to have some kind of structured and standardized representation of data. Thus, reducing significantly the amount of effort and time used to access this information and find relations and patterns, making the process of communication and application of this intelligence easier.
For this task, Ontologies are a great solution. By representing knowledge as a collection of classes, attributes, and relations between them, the ontologies could help an organisation to make better use of its data, granting a valuable advantage in a world governed by information. And with the swift expansion of AiaaS, handling of this data is becoming more and more vital.
Multi-Agents in AI as a Service
In artificial intelligence (and especially in AI-as-a-Service) you may have heard, very often, the term “Agent”. There are many types of agents, from the simplest ones having just “reactive” capabilities, to the most sophisticated ones having “reasoning” abilities, but they all share a common property: autonomy.
Before making some considerations about the growing AI-as-a-Service opportunity, we should clarify what’s the difference between the aforementioned “autonomy” and the older and simpler “automation”. Of course, a machine makes actions “automatically”, but the most interesting things happen when you are dealing with things so complex and unpredictable that you can’t program them directly one by one (the most known example probably is autonomous cars in a normal city environment). In these types of scenarios, you must find algorithms that give a reasonable output (the decisions of the agents) even if they face unknown situations.
As another example, think about one of the most common use cases of agents in software applications: customer service. How can you predict what customers will say? You will never have a list of all possible customer sentences or problems that they need to be solved: you need AI.
Now imagine you don’t have to study all of the difficult math needed to create the agents you need in your Company… here come the agents of AI-as-a-Service! A fast-growing opportunity for all types of industries and businesses. Keys in hand.
EsseSystems ECAS
The term complex adaptive systems are types of systems of systems that can adapt intelligently and generate new capabilities, functions and methods using cognitive and AI principles set out by the original founders of artificial intelligent systems. A human is a type of complex adaptive systems and with many biological, chemical, cognitive, electronic and physical systems orchestrated to form human intelligence, behaviours and functions.
The EsseSystems Complex Adaptive System (ECAS) platform does the heavy work of managing design, manufacturing, re-engineering and optimisation of products and services across , providing a low-code generative and intuitive environment to automate and adapt product and services, transforming them to a living organ, based on genetic algorithm, symbolic languages and multi-agent-based technologies, – so that you can focus on what matters the most: Building awesome products and services.
Recent Comments