Friday , February 22 2019
Home / chile / Two decades after the historic Deep Blue triumph, machines meet with people for the first time in the debate

Two decades after the historic Deep Blue triumph, machines meet with people for the first time in the debate



February 10, 1996 In Philadelphia, Deep Blue IBM made history by winning a chess match with world champion Gary Kasparov, showing the possibilities of data analysis I have not seen until then. History has traveled the world and has become a milestone in technological progress.

Despite fears, Terminator and Skynet style, which dominated the world, technology continued to develop, and today even mobile phones have a certain component of artificial intelligence (AI). And on Monday there was another chapter in the history of man versus the machine.

Exactly 23 years after Deep Blue vs. Kasparov, in San Francisco – the heart of new economics and technological progress – IBM presented itself for the first time in the public debate of Project Debater (PD), a machine created to understand and discuss problems with people. Unlike Deep Blue, this new computer is not just about "brute force", meaning the ability to process millions of alternatives to win an objective game like chess. The Project Debater must win in a more subtle field – arguments that must be coherent and convincing, and that's where AI is the key.

In the other corner was an expert on the championship debates, Harish Natarajan, who has a record of victories in this type of tournaments. It's kind of Kasparov, but it's a debate.

In a crowded auditorium, the presenter emphasized that what will be a witness will have a history, regardless of the result, especially since there are no objective winners or losers in the competition, but it depends on the jury, in this case from the viewers themselves . The topic to be discussed was known both by the computer and by Natarajan just 15 minutes before the debate, and it was "whether to subsidize the pre-school system or not". PD had to defend a position that he must successfully subsidize, and Natarajan rejects him.

In the liberal city and the US region the PD position was by far the most popular. The audience was forced to vote a few minutes before the debate with which they agreed, and 79% agreed that these posts should be subsidized. The first advantage – though circumstantial – for PD.

The computer opened fire IBM, with a woman's voice argued with a series of data that seemed to draw from Wikipedia and academic newspapers, pointing to the benefits of more children going to kindergarten, so it was important to subsidize them. In this first intervention, the PD did small jokes, provided information, and also tried to predict what would be his opponent's argument. They have 20 seconds left.

When Natarayan's time came, he refuted the PD argument, extending the emphasis on the alternative use of resources that would be allocated to these subsidies.

During the counter-argument and as the project manager pointed out later, it was known that PD not only listened to Natarajan's statements, but understood them in his background and tried to refute a new data bank and indicate subsidizing one thing did not mean stopping subsidizing others.

Natarajan insisted on his opinion, especially on the country's budget lines, and that research has shown that the majority of middle class people would send their children to kindergarten, even if there were no subsidies. In the final round, the PD could not take over all the arguments, but insisted on his initial idea, while the man focused on what apparently was one of the weaknesses of the machine: explore the different scenarios in more depth

In the final result – in which it was necessary to vote again – Natarajan raised 17 points, which he agreed with him (from 13% to 30%), while PD dropped from the same 17 points, to 62%. A man's victory, for now.

The IBM project manager stressed – in the middle of the annual Think event – that in this first public appearance of PD it was possible to predict that the machine is sometimes repetitive, but is able to understand and disprove it. As with all technological advances, the crucial moment on Monday was not the end of competition between people and machines (or people working on the development of artificial intelligence). Deep Blue already seems a distant chapter, and today we see how the computer exceeds man in calculations and strategies. In the next few years, we will probably not be surprised that the master of debates, political advisers and public decision makers – with a fashionable battle of ideas – connects to the wall or uses lithium batteries. Although it is not Skynet.


Source link