1983 WARGAMES and 2019 AI
I'm sure, in the 1980s, several studies on Artificial Intelligence (AI) and robotics were being done very seriously (Expert Systems, Industrial Robots, etc.). However, these technologies were featured as protagonists in films that stirred the imagination of youth at the time, and I was certainly, one of them :D
Not long ago, I saw one of these movies again and was surprised by the fiction-reality relationship between what I saw in these movies and what we know about AI today.
The movie WarGames, starring Matthew Broderick, shows a high school computer geek student who, looking to find new computer games, accidentally connects with a supercomputer that controls the US nuclear missile arsenal and innocently starts a game between US and Russia in the middle of the Cold War and could trigger World War III.
I, at the time, also a 12-year-old boy who was already programming at BASIC and was in the first steps of study in COBOL language, my eyes sparkled when I watched this movie. Having a microcomputer like the IMSAI 8080 (or MITS Altair) with blue phosphor monitor, 8-inch floppy disk drive, remote modem connection, 132-column printer was a dream with the full package.
But back to AI, despite the film showing other important points to date in the area of Information Technology such as Social Engineering, Geopositioning and Hacker culture, the WOPR (War Operation Plan Response) supercomputer BEHAVIOR - an IBM NA/FSQ-7, or simply Q7, the largest computer ever built - it was remarkable. And this is where I make the counterpoint today.
NLP - Natural Language Processing
In the scene where it accidentally accesses the WOPR supercomputer, the supercomputer establishes a terminal DIALOGUE that, only with reasonably sophisticated NLP algorithms would such interaction be possible.
ML - Machine Learning
In another scene of the movie, there is a presentation of the WOPR supercomputer (Q7 already shown above) to a government representative. At this point, it is clear that huge amounts of data are needed for the supercomputer learning process (it looks like they were already using Big Data and predicting Deep Learning).
We may think that the learning processes used by the WOPR supercomputer during the final scenes of the movie have been in two ways:
Supervised: In the scene showing Tic Tac Toe is played with the protagonist or having the machine play against itself, for the simplicity of the game, probably, the game had already been properly trained;
Reinforcement: In the later scene, when the missiles are launched, precisely because it is a dynamic simulation, the supercomputer evaluates each decision based on prizes and punishments, that is, the machine actually learns autonomously, which generates the tension over which result the machine would arrive.
Other AI Applications
Imagine what you can understand today about the Robinson family's Robot B9 BEHAVIOR of the 1965 Lost in Space serie?!! - or the technologies had seen in the cartoon The Jetsons created by Hanna-Barbera from 1962?!! - both reprized in Brazil in the 80's!
This is imagination and knowledge, hand in hand through time!
JSI:.