How Did We Come So Far? - Taking a Few Steps Back and Tracing AI's Journey
Posted by Tirthankar RayChaudhuri on Dec 05,2023
We are have now well and truly embarked upon the era of Enterprise Machine Intelligence Systems.
The AI industry today is already worth half a trillion dollars and is expected to exceed $2.0 trillion in 2030.
At this point I’d like to recount some of the major evolutionary milestones in the birth, growth and development of the field of Artificial Intelligence.
The below list is by no means an exhaustive compilation of all notable occurrences in the field of AI: such a list would be significantly longer.
Its worth noting that after 2010 the focus and emphasis of human endeavor has been on engineering advanced industrial solutions using machine learning/deep learning in preference to further advancing our scientific comprehension of the cognitive mechanisms of the human brain.
1943 The ‘Perceptron’ algorithm of supervised learning of binary classifiers developed by McCulloch and Pitts.
1948 Manchester Small Scale Experimental Engine built as the first known stored program computer able to execute a program stored within its electronic memory.
1950 Alan Turing publishes his paper ‘Computing Machinery and Intelligence’ proposing the Turing Test as a means of determining whether a machine can be classified as intelligent.
1955 Work commences on development of the First AI Program called the ‘Logic Theorist’ by Newell, Simon and Shaw. This program proved 38 out of the first 52 theorems in Whitehead and Russell’s Principia Mathematica using a search tree, heuristics and list processing.
1956 Conference organized at the Dartmouth College by John McCarthy with noted thinkers where the term ‘Artificial Intelligence’ was propounded.
1958 Frank Rosenblatt builds a machine to implement and demonstrate the capability of McCulloch and Pitts’ perceptron algorithm as the first officially recognized artificial neural network.
1965 Theory of Universal Grammar of languages propounded by Noam Chomsky.
1968 Release of Stanley Kubrick’s film 2001 A Space Odyssey inspired by Arthur C Clarke’s short story ‘The Sentinel’ (1951) which introduces the concept of an artificially intelligent supercomputer HAL 9000 capable of reasoning, speech, learning from experience, controlling a spacecraft, playing chess and expressing feelings.
1969 Publication of Minsky and Papert’s book on Perceptrons proposing a shift in AI theory from symbolic reasoning to connectionist systems based on the earlier work of McCulloch and Pitts and Rosenblatt.
1973 Publication of the Lighthill report by James Lighthill for the British Science Research Council with highly negative and pessimistic criticism of the field of artificial intelligence.
As a consequence of The Lighthill Report funding of research to the field was severely impacted in the 1970s and 80s.
1986 Ground-breaking paper on ‘Learning representations by back-propagating errors’ published in the journal Nature by David Rumelhart, Geoffrey Hinton and Ronald Williams.
This publication is credited as being the birth event of deep learning as we know it today.
1991 Python programming language developed by Guido van Rossum.
The majority of machine learning programs today are written in Python.
1997 IBM’s Deep Blue supercomputer defeats world chess champion Gary Kasparov in a 6-game match.
2005 Jeff Hawkins founds the company and product Numenta based on a cohesive AI model called hierarchical temporal memory (HTM) which emulates layered reasoning levels within the human brain’s neocortex.
2011 IBM’s AI computer Watson wins the trivia game show Jeopardy defeating the two all-time topmost champions of the game.
2012 Google Voice Search (using deep learning) launched having a mere 8% error rate in recognizing spoken words thus exceeding average human capability.
2014 Generative Adversarial Networks (GANs) published by Ian Goodfellow as a major breakthrough in image generation technology
2015 Tensorflow AI development platform launched by the Google Brain team.
2016 Pytorch machine learning framework launched by Meta AI (earlier Facebook).
2017 Transformer deep learning for Natural Language Processing and Computer Vision launched by Google Brain.
2022 OpenAI launches their advanced chatbot Chat GPT which uses the Transformer algorithm to generate a large language model (LLM) capable of reinforcement learning.
2023 Generative AI emerges as disruptive force with the potential to transform the way businesses operate.