Intelligence is the ability to adapt to change – Stephen Hawking

Intelligence is the capacity to learn and solve problems – (Webstar’s dictionary)

The field of Artificial Intelligence endeavors to develop humanlike ‘intelligence’ in ‘dumb’ computers. Humans have an inborn capacity to learn, absorb and behave rationally in certain environments and situations, and it will be difficult for computers to replicate such desirable behaviors in real-life situations. That is why, till very recently, at least till the advent of the subject Artificial Intelligence and its developments, computers were considered ‘dumb.’ Can a computer recognize a person from his changed features in a recent image when the computer has an image of the same person taken years back in the database?

The field of AI builds intelligence in computers to duplicate human performance in certain situations. When your computer matches your skill or even beats you in a game of chess, you’re actually facing an opponent that is capable of interpreting your moves and come up with a better one to have you checkmated. Wouldn’t you call such an entity, virtual it may be, intelligent? The chess software is an example of an intelligent program.

The term ‘intelligence’ has befuddled computer scientists since the inception of the subject Artificial Intelligence. The early researchers of AI claimed that to be branded ‘intelligent’ the entity must be self-aware of its existence and model imaginative responses to external stimuli from the Universe. The adherent to this school of thought negated the question that arose in the AI community ‘Can a machine think?’

It was only in the 1970s that Alan Newell and Herbert Simon incorporated the idea of ‘intelligence’ in machines. In their hypothesis called ‘Physical Symbol System hypothesis’ they assert that the ability to represent and manipulate symbolic structures is both necessary and sufficient to create intelligence (Newell and Simon, 1976). This means that it does not matter whether the agent is living or non-living. The requirement is that of an ability to create symbolic representation & manipulate those representations. The degree of intelligence is dependent upon the fitness of the representation and the effectiveness of the manipulation procedure [1].

In their book, Russell & Norvig [2] highlights the following as essentials of intelligence: The ability to solve novel problems, the ability to act rationally & the ability to act like humans.

AI experts concur that Intelligence requires Imagination and provides an estimate of our ability to comprehend meaningfully the changing world – our capability to sense and reason suitably the factors that has the potential to impact us and initiate subsequent course of action.

Learning is an essential ingredient for Intelligence. Learning through Explanation, Observation, Rote or Experience and various other ways can evolve into knowledge, to be kept in knowledge base, and made use of for manifestation of intelligence in a world where an increasingly complex decision making scenario greets the Agent.

Can AI rival human intelligence?

As of now, humans outclass computers in Face Recognition and many other Pattern Recognition tasks, the deductive and processing ability working out results in a jiffy in the human reasoning system. Computers take much longer using algorithms and matching data values.

Desirable pointers towards Intelligence:

• Observational power (Pattern Recognition)
• Fast Learning
• Predictive ability (Forecasting)
• Intelligent guess (Heuristic)
• Adaptability (incorporate changes)
• Decision making ability under Uncertainty (Fuzzy Inference)
• Ability to recall/find similarity of past situations and make use of experiential knowledge (Case Based Reasoning)
• Etcetera

References:

1. Deepak Khemani, A first course in AI, McGraw Hill Education
2. Stuart Russell & Peter Norvig, Artificial Intelligence: A modern approach, ISBN-13: 978-0136042594