How exactly do you define Artificial Intelligence(AI)?
This looks like a back to basics/ back to school question – but the answer is not that simple
Recently I was trying to find a good academic definition of AI for a research paper.
Surprisingly, its not easy.
In this post, I present a good definition for AI from an academic standpoint and also explain the rationale behind that definition
There are two streams of thought when it comes to AI: On one hand, we have the data-driven approach based on neural networks and deep learning. On the other hand, we have the relationship of AI with human thinking and cognitive abilities.
To make it more interesting, over time, the emphasis on these approaches has shifted.
The original (1955/1957) Dartmouth Summer Research Project on Artificial Intelligence1 aimed to “make machines use language, form abstractions and concepts, solve kinds of problems now reserved for humans, and improve themselves.”
This definition/ description tends to the human / cognitive aspects of AI and was initially implemented by methods based on the manipulation of symbols, sometimes referred to as ‘good old-fashioned artificial intelligence (GOFAI)’.
In contrast, neuroscience-inspired approaches such as connectionism and neural networks focused on learning and representations, which led to the data-driven deep learning methods that are prominent today.
Indeed, many people now consider AI to be synonymous with deep learning.
However, deep learning in its current state has limitations: it tends to be data hungry, compute heavy, sensory-based, prone to unexpected mistakes as shown by adversarial examples, and inefficacious for much of cognition and behaviour.
Ultimately, the GOFAI approach was overtaken by neural networks. But today, we see an increased discussion on hybrid methods such as the use of AI to play Bridge(a game that needs reasoning and planning which is not easily possible in conventional neural networks)
Thus, the meaning of AI will remain in flux – making an exact definition hard.
So, the best definition I could find is in an article in journal nature
“Artificial intelligence can be defined as intelligence demonstrated by machines. But what counts as intelligence, and how intelligence is implemented in different kinds of machines, robots and software varies across disciplines and over time.”source https://www.nature.com/articles/s42256-020-00262-2
Source nature – Materializing Artificial Intelligence
Image source pixabay