Return of the 10th Circle Archmage

“Alphagon won the match against Lee Sedol 9-dan in 2016. How much notation did Alphagon really need?”

“·················ogue.”

The students were silent again at Professor Haeun Yoo's question.

‘30 million chakjeom information and 160,000 professional Go players’ notations.’

I replied eagerly to myself.

“Is this too difficult a question? Or are you afraid of being wrong?”

Asking such a question, Professor Yoo Ha-eun spit out the same answer I had come up with.

“Anyway, not only did Alphagon go through this data learning process, but in the actual Go match against Lee Se-dong, about 1202 CPUs and 176 GPUs connected to the cloud were operated simultaneously, and through about 340 million repetitions during the 5th round. We won.”

“340 million times!!!”

A total of 361 spaces.

The number of cases in which the first number is exchanged, 129,960.

There are 10^170 ways to fill in all 361 points.

In 'Baduk', a field that even humans have not yet conquered.

It's really difficult to calculate notation even if it's just one plate per person, Alphagon alone... It was.

“Is this just a surprise?”

“·················ogue.”

"how is it? Alphagon is so smart, what were you guys thinking right now?”

“·················ogue.”

Professor Yoo Ha-eun clicked his tongue as if he knew everything without seeing it.

“After class, you are probably wondering what the school menu is today, right? Am I wrong?”

“·················ogue.”

Actually I was just thinking about it.

“Artificial intelligence is not limited to Alphagon. Alphagon is just the beginning.”

In fact, it could be seen that artificial intelligence beating people in Go was simply gossip.

What really matters is...

“Where will such smart AI be used in the future? Soccer? baseball? Or a game?”

As Professor Haeun Yoo said, a lot of artificial intelligence has already been introduced and used in games.

'Because games are used a lot for machine learning of artificial intelligence...'

Is it a simulation?

Artificial intelligence was not just about designing algorithms.

In terms of living things, the algorithm was like a newborn baby, and we had to learn the algorithm through iterative learning through big data.

‘Machine Learning, Deep Learning, Reinforcement Learning’

In the meantime, there have been various learning processes in artificial intelligence.

At first, humans wrote the coding suitable for the artificial intelligence algorithm and learned it manually one by one.

However.

‘Now is the time for artificial intelligence to learn on its own by using visual and auditory functions.’

As a result, the development speed for artificial intelligence has increased dramatically.

'Because the ability to see and hear has already surpassed that of humans...'

Artificial intelligence did a better job of accurately discriminating what it saw and heard through scientific experiments.

ImageNet7 contest where various research institutes participate every year to compete for the accuracy of object recognition in images.

In 2022 there, Nanosoft achieved an accuracy of 97.85%, surpassing the recognition rate of humans (94.90%).

Complex reasoning was not yet possible, but in human terms, it caught up with the complex reasoning of a 7-8 year old human.

“Artificial intelligence is pervasive in our daily life.”

Starting with smartphones, drones for delivery, self-driving cars that are just beginning to emerge, etc...

Professor Lee Ha-eun explained about the ripple effect on our society due to the development of artificial intelligence in the future.

“Not long ago, an analyst analyzed the impact of artificial intelligence on jobs in Korea.”

Then I picked up the board mark and started writing the results on the whiteboard.

“As a result of the analysis, it was analyzed that 43% of all jobs in Korea will be replaced by artificial intelligence in the future.”

“43%………….”

At 43%, it was close to half of all jobs in Korea.

“Then, which of those job groups will disappear first?”

Whether it was a question asked for an answer, Yoo Ha-eun immediately wrote the answer on the whiteboard.

“By occupation, it is about 70% for office workers, sales workers, and machine operation workers, and by industry, about 60% for jobs in wholesale and retail, food and lodging, manufacturing, etc.”

“·················ogue.”

The students were so startled that they were now at a loss for words.

'I knew that a lot would disappear, but seeing it as a percentage like that, it's new again...'

Again, there was a difference of dozens of times between thinking roughly and accurately recognizing numbers.

“By educational background and income, the high-risk share is highest in certain classes of jobs. So who will be hit the hardest?”

‘Middle class.’

I replied hard to myself again.

‘In the industrial revolution, the ambiguous middle class is always swept away first.’

It was the same in the third industrial revolution.

Originally, in the 80s and 90s, the office was occupied by typewriters, stenographers, and abacus technicians.

It was gradually replaced by computers armed with Word and the Internet.

As various office simplifications using computers were made, various other office jobs were swept away like a tsunami.

'Now the 4th industrial revolution has come...'

Although computers have completely swept away the existing analog jobs, a myriad of new jobs related to computers have emerged.

‘At the same time, existing industries have also achieved office efficiency ...’

In the end, office workers had to obtain a Word or Excel certificate as a must to survive in the workplace, and the era of computer qualifications that speaks of a person's abilities has come.

'If computers were to help humans in their weak areas of computation...'

Complicated mathematical calculations were too inefficient to do with the brain. Because the human brain wasn't made to be used that way.

'But AI is different...'

As I expected, Professor Haeun Yoo began to explain the algorithms of artificial intelligence and its ripple effects.

“The ‘Polanyi paradox’ has been suggested as a basis for the difficulty of automation of unstructured tasks by artificial intelligence.”

Polanyi's Paradox was the hypothesis that 'people know more than words can say'.

Explicit rules that are hard to put into words.

That meant it was difficult to learn with a computer. It was virtually impossible to follow the tasks of learning through experience, such as driving a car, and responding by judgment according to the situation with a computer program.

However.

“Current AI technology is pushing the boundaries of what can be automated by going beyond the Polanyi paradox.”

Professor Yoo Ha-eun vomited hard, as if explaining the birth of a new 'bell' that is taking place with a pounding heart.

“Now, AI itself learns data through trial and error and improves errors through algorithms. Just like humans, it has become possible to acquire knowledge through experience.”

machine learning.

In other words, machine learning meant self-learning from data for actions not specified in code.

“And among machine learning methodologies, deep learning has recently been in the spotlight.”

deep learning.

In other words, deep learning, based on a deep neural network, was an artificial intelligence implementation method that mimics the operation of the human brain.

“Deep learning has allowed us to simplify or skip the cumbersome and costly process of data manipulation.”

In such a way, deep learning has produced very superior results compared to past algorithms, such as pattern recognition and natural language processing.

‘Even in the case of Alpha Gon, who defeated Se-dong Lee, it wasn’t the way to input the way to play Go well into the computer. Because it was implemented so that the computer learns the game data of the Go masters and learns the winning strategy by itself....'

Alphagon, the representative of deep learning.

After accumulating basic skills in the notation of numerous Go knights, Alpha Gon did Reinforcement Learning, playing against himself millions of times. So he was able to accumulate skills that surpassed human masters.

“As the supply of remote unmanned stands is already increasing, jobs in the service industry are disappearing around us, aren’t we?”

It really was.

Compared to developed countries such as the United States, Japan, and China, Korea has been far behind the 4th industrial revolution.

However, even in such a country, the winds of change were blowing.

‘When Hye-eun went to Randaria a while ago, she placed an order in front of the unmanned checkout counter . . .In the past, an employee would stand at the checkout counter to take orders from customers, and the 'short lathella' to the cooking staff behind the back has changed.

‘Almost everything is paid by card instead of cash, and orders are made without going through an employee using the unmanned touch register…’

So the person who took the order disappeared, and now only the person who cooks inside and the person who wraps it up remain.

‘Other industries will lose their jobs one by one.’

As I expected, Professor Yoo Ha-eun added confidence to my thoughts by entering overseas cases.

“Yamazon is already offering drone delivery, unmanned stores, and unmanned delivery services through Yamazon Go.”

That's why online and offline stores are either going bankrupt or downsizing their business by being pushed by Yamazon.

Yamazon, a global leader in artificial intelligence technology, continued to expand its influence and dominate the distribution industry through continuous innovation such as product recommendation using artificial intelligence and warehouse automation using robots.

“And in China, ‘BingoBox’, a 24-hour unmanned convenience store, ‘Tao Cafe’ of Alibaba, ‘TakeGo’ of food company Wahaha, etc…”

The thing that cannot be ignored now is that China was far ahead of us in the 4th industrial revolution.

*

"Huh, brother... I’m so scared I’m going crazy.”

As soon as the lecture was over, Lee Ji-eun sighed and I complained.

“What are you afraid of?”

i'm having fun

Swallowing the back words, I asked Lee Ji-eun.

“Artificial intelligence is taking all our jobs…. Whoops..."

“·················ogue.”

I almost asked, 'What have you heard so far?'

Feeling that way, Lee Ji-eun worked hard to express her thoughts.

“Because my father actually drives a truck……”

"Well······."

Lee Ji-eun said her father was in the long-distance transportation business and was responsible for the livelihood of his family.

“But, what if a self-driving AI takes my father’s job?”

I nodded at Lee Ji-eun's complaint.

It was quite probable.

‘The transportation industry is an occupation that is included in 72% of the high-risk jobs that will disappear in the future.’

I thought so, and said to Lee Ji-eun as if it wasn't a big deal.

“Then you can get a job and take responsibility for your family. Not like anyone else.”

"Yes???"

At Lee Ji-eun's dazed reply, I burst into laughter.

‘Who is the group that will dominate the business with artificial intelligence in the future?

It couldn't have been a really useless military worry.

Thank you for reading this story at mtlarchive.com. Your support enables us to keep the site running!