IIT_CS480_Assignments_Week4_Participation

IIT CS 480 - Introduction to Artificial Intelligence
Assignments_Week4_Participation

github: https://github.com/IIT-MAI-2020Fall

---

This week, we will discuss Chapter 5 - Games. We’ll cover mini-max, alpha-beta pruning, and briefly about Monte Carlo Tree Search. This discussion is about Deep Blue and AlphaGo.

Deep Blue

Read: https://en.wikipedia.org/wiki/Deep_Blue_(chess_computer)

Watch: https://www.youtube.com/watch?v=5hRNlfAUeEE

AlphaGo

Read: https://en.wikipedia.org/wiki/AlphaGo

Watch: At least the first 13 minutes of https://www.youtube.com/watch?v=WXuK6gekU1Y

Discuss

Deep Blue
Do you think Deep Blue thinks like human? Why or why not?
Do you think Deep Blue acts like human? Why or why not?
AlphaGo
Do you think AlphaGo thinks like human? Why or why not?
Do you think AlphaGo acts like human? Why or why not?
What do you think the next game challenge for AI is? Or do you think the games are already a solved problem for AI? Please elaborate.
(Optional) If you have reading, movie, or short video suggestions regarding Deep Blue, AlphaGo, or some other AI system playing games, please share.

---

1. Deep Blue

1.1 Do you think Deep Blue thinks like human? Why or why not?

1.2 Do you think Deep Blue acts like human? Why or why not?

1.1 No. “Deep Blue used custom VLSI chips to execute the alpha-beta search algorithm in parallel.It was a brute-force search approach.”（quoted from Wiki). So as hardware advances in capability, it’s only a matter of time before that machine beats the human brain.

1.2 As far as chess is concerned, yes. Because the player can’t tell whether the other side is a machine or a human.“After the loss, Kasparov said that he sometimes saw deep intelligence and creativity in the machine’s moves, suggesting that during the second game, human chess players had intervened on behalf of the machine, which would be a violation of the rules.” I don’t know much about chess, and I don’t know how the “deep intelligence and creativity” in chess is expressed.But I think that since the performance of chess is only “put a piece somewhere”, then people should be indistinguishable. So, it’s ok to say Deep Blue acts like human.

2. AlphaGo

2.1 Do you think AlphaGo thinks like human? Why or why not?

2.2 Do you think AlphaGo acts like human? Why or why not?

2.1 No,it isn’t. According to the wiki, AlphaGo is based on a Monte Carlo tree search(MCTS). But the way humans perceive things is much more complicated than that.

2.2 Yes. According to principle of operation in MCTS, "The final game result of each playout is then used to weight the nodes in the game tree so that better nodes are more likely to be chosen in future playouts."The behavior of constantly learning from failures and correcting them is very similar to the human learning process.

3. What do you think the next game challenge for AI is? Or do you think the games are already a solved problem for AI? Please elaborate.

In my opinion, AI can gradually replace jobs with clear boundaries of input and output. For random work of input and output, it is a great challenge to give a reasonable answer.Such as psychological counseling, artistic creation, children’s teaching.

In aspact of games, whether AI can solve the game problem depends on the classification. If it is a competitive game, AI can be competent; if it is a sandbox game, I think AI will not know what to do, such as Minecraft, GTA5, etc.

4. (Optional) If you have reading, movie, or short video suggestions regarding Deep Blue, AlphaGo, or some other AI system playing games, please share。

I recommend westworld, which tells the story of a huge high-tech adult paradise themed with Westworld, providing tourists with the satisfaction of killing and sexual desire. With the independent consciousness and thinking of the agent, they begin to doubt the nature of the world, and then wake up and fight against human beings.

In addition to the technical elements, the ethical issues raised in the show about artificial intelligence and technology are also the questions we may face in the future.

---

欢迎 勘误 & 分享 & 交流