Robots. Self-driving cars. Emotion and Image Recognition. Smart cities.
There has been quite very interesting developments in the field of technology and computer science during the previous years. One key factor that drives the success of the advanced technologies people are working on are based on Artificial Intelligence (AI).
For the past decades, humans have created machines that can do specific and pre-determined things automatically or with the control of human hands and minds. Now, as humanity has evolved, learned and discovered several aspects of reality, there is now a race in understanding human intelligence through AI by creating something that would somehow imitate how humans act, react or decide on things. To put in simpler terms, artificial intelligence can make a machine decide for itself.
We were asked to watch the documentary of DeepMind, an AI company and research laboratory, on how they created a program about Go, a complex Chinese board game which started over 3,000 years ago. Creating a program on this board game is different from creating a program to play Chess. For Go, there are around 200 different moves for each turn that can lead to numerous outcomes that would require heavy computing power while Chess will just need fewer than 50 moves.
DeepMind computer scientists and programmers had trained different datasets of possible moves from amateur and professional players of the board game, incorporated a probability and tree search for the next following moves which has led to creating AlphaGo, a machine that can play Go with great predictive capabilities.
To test the power of AlphaGo, the top player in the field worldwide, Lee Sedol, has been invited to play the game with DeepMind’s creation. Lee and his supporters were positive that he will win against AlphaGo, but to everyone’s shock, he was defeated even on the first round.
Using deep learning and neural networks, DeepMind has created a powerful program that can dominate even the top world players who dedicated their lives into playing Go. Because of its computational power and trained datasets, AlphaGo transcended how humans play the game Go.
I honestly didn’t know about this game until I watched the documentary and although I didn’t quite know a lot of details on the game, I was quite hooked with what I was watching. What made everyone baffled was Move 37 of Round 3 which showed that AlphaGo has learned, developed and proved that it can make a creative move intuitively like how humans do. Move 37 was not one of the top 5 choices that professional Go players will make. It was also stated that move 37 is 1 in 10,000 chances to be done by humans and AlphaGo surprised Lee and even the developers of the program with this move.
Although AlphaGo was always steps ahead of Lee Sedol, Lee has also shown that even though AI has limitless powers, it still has its weakness points and vulnerabilities that can lead to its downfall. Lee’s Move 78 of Round 4 hit AlphaGo’s deep learning capability which eventually led to several questionable moves and in the end, its resignation and Lee’s win. His win during Round 4 was like a win for humanity and I was actually cheering for Lee, alongside with other of his supporters.
Creating AlphaGo was a feat in itself in the field of Artificial Intelligence. But there are still a lot to discover, develop and to be worked on. So far, as for Kai Fu Lee, a former big tech executive, China, coupled with their competitive culture, government backup and vision, top minds and focused implementation, may have been leading in the race of AI compared to the rest of the world. As the Western world always has new breakthroughs regarding AI, China on the other hand, focuses on improving and implementing current AI tech using the datasets that they have. AI will not be successful without data and with China’s current policies and digital environment, massive data is gathered second per second in all the actions that Chinese people do.
Yes, data is the new oil and it is very important in the progress and development of several technologies regarding AI but data ethics and data privacy is also another thing to look at especially when harvesting information from people.
It is amazing but at the same time scary to realize what AI can do once it reaches its peak, and it is also interesting to know how AI’s peak will look like. Given that humans make more unique and original moves and humans in general, are very much complex, I also wonder when and how AI will truly match human intelligence.
AI will be really helpful in improving and developing society’s processes and solving our current world’s problems but similar to other technology and inventions, under the wrong hands and if not used wisely, this may lead to one of the solutions-turned-problems that again needs to be solved. We don’t want to have machines or terminator-like robots ruling our planet and if this would happen, humans will need a God-like touch and resistance movement such as Move 78.