Artificial Intelligence (AI) has been the talk of the town these days. From hyperlocal startups to Multinational companies, everyone has either started using AI in their businesses or are having plans to begin in the quarters to come. It wouldn’t be wrong to say:
“dot-ai is the new dot-com.”
- Jash Rathod
In my previous post, we explored, in a general sense, what AI is. If you haven’t explored it yet, feel free to check it out by clicking on the following link: Simplifying Artificial Intelligence.
In this post, we will dive deeper into AI and explore its various categories. Don’t worry if this sounds overwhelming; you’ll be fine! If you are a student, someone who is or aspires to be an AI engineer/researcher someday, a business owner or an entrepreneur, or someone for whom perhaps Computer Science is a new subject altogether, this series is for you. Mark Cuban discusses the impact of AI as:
“If any of you are entrepreneurs or in the business world and if you don’t know AI, then you’re the equivalent of somebody in 1999 saying, ‘Yeah, I’m sure this internet thing will be okay, but I don’t give a sh*t.’ It’s the same thing.”
- Mark Cuban, American Entrepreneur, Television Personality, and Investor.
Well, I believe, regardless of one’s field, everyone must at least have a basic understanding of AI. Who knows, maybe, the magic of AI helps you take your domain of work to the next level!
Let’s begin by defining AI, as discussed in this paper.
“A system’s ability to correctly interpret external data, to learn from such data, and to use those learnings to achieve specific goals and tasks through flexible adaptation.”
We humans learn from experiences and so does AI. These experiences need to be fed to the machine as values (like numbers, text, images, etc.) When a machine reads this input, it tries to make sense of it by employing the use of mathematics. It defines an equation or creates a storage mechanism like vectors or matrices, each of which serves as a “brain with stored knowledge” to the machine. In the discussions further, we will refer to this mechanism as a ‘model’. So, an AI model takes data as input, learns from it, and acts in accordance with its learnings.
AI can be categorized based on its Capabilities as:
Artificial Narrow Intelligence (ANI)
ANI or sometimes also called ‘Weak AI’ is the case when our model is highly skilled at performing a single task or sometimes, a handful of tasks. These models can be as good as humans at those tasks or even better than humans. Most of the work done in the domain of AI has been in ANI.
Examples:
- Self-Driving Cars
- Pet Classifiers from images (Dog vs Cat)
- Essay Grading Systems (Yes, AI can do that!)
AI can be a better classifier of pets given their images or could be a better essay grader than humans, but a model developed for a pet classifier cannot be used for grading an essay or driving a car; thus being “narrow” in scope.
Artificial General Intelligence (AGI)
AGI or ‘Strong AI’ is the concept of a machine with general intelligence that mimics human intelligence and/or behaviors, with the ability to learn and apply its intelligence to solve any problem. It possesses the ability to think and act in a way not any different from the way we humans think or act. We humans can drive a car, classify a pet by looking at its image, and grade an essay all with just one brain of ours. While an ANI system cannot do so, an AGI can! True AGI has not been achieved yet by AI researchers and scientists, owing to intricacies and complexities that it proposes. Who knows, maybe someday technological advances like Quantum Computers can enable achieve True AGI!
Artificial Superintelligence (ASI)
ASI, as the name suggests, is where the machines surpass human intelligence and ability. Robots destroying the world? — that’s ASI. AI is far (believe me, really far …) from achieving this stage. If someday, it was to become a reality, the decision-making and problem-solving capabilities of the super-intelligent beings would be far superior to those of human beings, and its consequences, beyond our imagination.
AI can also be classified based on its Functionalities as follows:
Reactive Machines
This category is like a human’s response to stimuli. They are the oldest form of AI systems and do not have a memory-based functionality. These are suitable for tasks that require actions based on just current situations and not past experiences. Eg: Chess-playing machine.
Limited Memory
Where Reactive Machines lack the memory element, these are memory-based, along with all the functionality that the former offers. As these possess a memory to store past experiences, they can learn effectively and suggest a more calculated solution. Most of the applications that we see today fall under either the reactive machines or the limited memory category. Eg: Apple’s Siri.
Theory of Mind
This sector is a budding one and research is still needed for it to be classified as complete. It is expected from this type of AI that it should be able to understand human emotions, beliefs, thoughts, and the ability to interact socially — something of a conundrum for the heartless and apathetic piece of metal.
Self-Aware
This is the end goal for AI development. It currently exists only hypothetically. It is an AI that has evolved to be so akin to the human brain that it has developed self-awareness. Creating this type of AI is decades, if not centuries away from materializing. This type of AI will not only be able to understand and evoke emotions in those it interacts with, but also have emotions, needs, beliefs, and potentially desires of its own. And this is the type of AI — the Ultron — that doomsayers of the technology are wary of.
From what we discussed above, your Sherlock mind would have started to contemplate over how AI can solve some problems in your domain, your business for you and your customers, or some ubiquitous problem hidden in plain sight. Moreover, you will now be better equipped to tackle and decide what kind of AI can help you solve it.
This is all for today.
Gracias!
Have an amazing idea that you would like to discuss with me? Or wish to share your thoughts on AI? Worry not, feel free to get in touch with me on LinkedIn. Always happy to help!