2021年12月4日土曜日

Kobe's small-scale IT president who challenges the development of the Nobel Prize-class "heart for robots"

https://news.mixi.jp/view_news.pl?id=6764072&media_id=262

I shared it.

 

December 03, 2021 20:30   Maid News

Mysterious news

PhotoAtsushi Takata.  Since the establishment of the company in 2009, most of the profits obtained from the products developed in-house have been devoted to the research and development of "programming the mechanism of the mind".
Atsushi Takata. Since the establishment of the company in 2009, most of the profits obtained from the products developed in-house have been devoted to the research and development of "programming the mechanism of the mind".

Looking at a robot that behaves like a human, it's no surprise now that it's a pre-prepared production. Only in the world of manga and anime, robots themselves have emotions and thoughts. Even if you can talk to people, you can't have a robot with a heart. Most people think so.

[Image] A program of mental mechanism is indispensable for robots to have a natural conversation with humans.

With deep learning (a technology that discovers the characteristics of data from a large amount of data), which is currently the mainstream of AI, robots that have natural conversations cannot be forever, so we are advancing AI development with a unique approach that programs the mechanism of the mind. Atsushi Takata (52), the representative of Robomind Inc. (Kobe City), is doing this.

One of the reasons why Softbank's Pepper, the world's first emotion recognition personal robot that appeared in 2014 with noise, did not spread is that it is "human" due to the planned harmonious conversation patterns and tongue-in-cheek answers that are clearly different from real humans. Takata says it was because he couldn't feel it.

Even if you let Pepper learn a large amount of conversation data, you cannot have a natural conversation. The reason is that conversations do not have clear rules like chess and shogi, and cannot be algorithmized (ruled). Then, if the minds (emotions and thoughts) of the person who controls the conversation are made into rules and incorporated into the program, a robot that holds emotions and thinks like a person will be born, and naturally conversation will be possible. Mr. Takata thought that what was important was not the amount of information, but the "mechanism for recognizing information" that was common to people.

The current AI robot is a simple logic that returns a sentence related to a word in a conversation. For example, even if an elementary school child returns home from school and happily talks to the AI ​​robot, "I got a pudding for lunch," the only reply is, "School lunch started in 1902." Can not. Mr. Takata is aiming for an AI robot that can understand the emotions of the other party and respond with the feelings of the other party, saying, "I'm glad I got a pudding."

That said, there is no way we can think in common sense, such as creating rules (mechanisms) for the mind. However, Mr. Takata collected and examined information from various fields and continued to publish his own hypothesis on blogs and SNS. Then, the number of people who were interested gradually increased, and excellent engineers who wanted to work for Mr. Takata gathered. The YouTube video that started two years ago now exceeds 200, and in addition to explaining the mechanism of the mind, the contents of the program, which can be said to be confidential, are also open to the public. And this year, a program that can finally reproduce the human heart, the name of which is "Mind Engine", has been patented. It was sunny and I was able to confirm that there were no similar development cases.

Mr. Takata develops his theory that we are not recognizing the real world as it is, but recognizing the virtual world created by consciousness based on the information obtained by the five senses. Mr. Takata named this the "virtual world hypothesis of consciousness," and this idea plays a key role in creating the mechanism of the mind.

Who is Mr. Takata who thinks that there is no such absurdity?

After graduating from the Faculty of Engineering of a national university, Mr. Takata will get a job at a patent office in Osaka City. In his late twenties, who was living a smooth sailing salaryman life, Mr. Takata suffered from a serious illness and was told that he had a "5-year survival rate of 30%." But mysteriously, I wasn't depressed. After seriously thinking "let's do what I really want to do", I arrived at "making a mind with a computer" and started learning programming from scratch. For the next 25 years, he has been energetically practicing the way of life he originally vowed.

It is not because of financial success or self-revealing desire that we can continue to challenge the Nobel Prize-class for more than 20 years. It's purely fun, doing what you want to do, that's it. However, recently, something like a sense of mission has sprung up, and the desire to complete a robot with a heart has increased further. If anyone is doing similar research, please let me know. If your idea is wrong, please refute it and correct it. Mr. Takata has such expectations for this article.

(Maidona News Special Contract, Moriyasu Kitamura)

0 コメント:

コメントを投稿