The Machine That Changed the World: Episode 4 - The Thinking Machine


1950 | 1960 | 1970 | 1980 | IA | MINSKY | Ordenador | SHANNON | TURING
Personaje: MINSKY, Marvin.
Personaje: SHANNON, Claude Edwood.
Personaje: TURING, Alan Mathison.

The Machine That Changed the World is a 1992 documentary series on the history of electronic digital computers, from the dawn of the computer in the 1800s to the early 1990s. It was produced by WGBH Television in Boston MA, in cooperation with the British Broadcasting Corp. (BBC), with support from ACM (Association for Computing Machinery), NSF (National Sciencie Foundation) and UNISYS. The series consists of five episodes.

Episode 4 - The Thinking Machine.

This episode covers the history of artificial intelligence and the challenges that come from trying to teach computers to think and learn like us.

The history of artificial intelligence, from Minsky to neural networks.

Key People/Concepts:
Jerome B. Wiesner, Marvin Minsky, John McCarthy, Oliver Selfridge of Lincoln Labs, Claude Shannon, Freddy Robot at the University of Edinburgh, Sir James Lighthill, Terry Winograd’s SHRDLU, Edward Feigenbaum’s work on expert systems, Doug Lenat’s Cyc project, Oliver Sacks, neural networks, NETtalk.

Marvin Minsky (MIT, died 2016),
Hubert Dreyfus (UC Berkeley, died 2017),
Edward Feigenbaum (Stanford University),
Hans Moravec (Carnegie Mellon Robotics Institute),
Doug Lenat (University of Texas, Austin),
Dean Pomerleau (Carnegie Mellon Robotics Institute),
Terrence Sejnowski (Salk Institute).

In this episode:
Late 1950s - Marvin Minsky and John McCarthy set up an A.I. Dept. at MIT.
1960 - Slagel's program for freshman calculus; from "number crunching" to intelligent problem solving.
Mind vs. Brain approach; mind = software, brain = hardware.
Block stacking program - lack of "common sense".
1970 - Edinburgh University, "Freddie" image recognition application.
1970s - Stanford Kart; motion planning.
Alan M. Turing.
Joseph Weizenbaum's ELIZA.
Russian to English language translator - earliest of the non-numerical applications.
Underestimation of the difficulty of A.I.
Future of A.I. looks bleak - Dreyfus' "What Computers Can't Do".
Terry Winograd's SHRDLU - intelligence within microworlds.
Expert Systems - Feigenbaum's DENDRAL.
"Idiot savants".
Early 1970s - story understanding via scripts and frames. (Minsky)
Modeling commonsense.
1984 - Lenat's ten-year CYC project to catalogue "commonsense".
A new look at modeling intelligence by modeling the biological brain.
1950s and 1960s - Perceptrons.
Late 1970s - Neural networks; Connectionists.
Self-driven vehicle - "trained" to drive.
Selective training - tank recognition failure.
Large networks require large training times.
Brain - a collection of special purpose machines --> general intelligence/ commonsense.
A.I. - history of fascinating failures.

Más en

Más en

Foro Histórico

de las Telecomunicaciones


c/ Almagro 2. 1º Izq. 28010. Madrid
Teléfono 91 391 10 66
c/ General Arrando, 38. 28010. Madrid
Teléfono 91 308 16 66

Copyright 2019 Foro Histórico de las Telecomunicaciones