Robotics

Robot hand wins rock, paper, scissors every time

Robot hand wins rock, paper, scissors every time
A robot hand developed by the University of Tokyo's Ishikawa Oku Lab is reportedly so adept at the game rock, paper, scissors that it is unbeatable against a human opponent
A robot hand developed by the University of Tokyo's Ishikawa Oku Lab is reportedly so adept at the game rock, paper, scissors that it is unbeatable against a human opponent
View 4 Images
Combined with the clearly very highly responsive actuators in the robot hand and wrist, the robot is able to throw its shape before the human player has finished their move
1/4
Combined with the clearly very highly responsive actuators in the robot hand and wrist, the robot is able to throw its shape before the human player has finished their move
Strictly speaking that should be robot hand and eye, since, as you might have guessed, it uses a high speed vision system that identifies, within a millisecond, the move being made by the human player from the shape of their hand
2/4
Strictly speaking that should be robot hand and eye, since, as you might have guessed, it uses a high speed vision system that identifies, within a millisecond, the move being made by the human player from the shape of their hand
Strictly speaking that should be robot hand and eye, since, as you might have guessed, it uses a high speed vision system that identifies, within a millisecond, the move being made by the human player from the shape of their hand
3/4
Strictly speaking that should be robot hand and eye, since, as you might have guessed, it uses a high speed vision system that identifies, within a millisecond, the move being made by the human player from the shape of their hand
A robot hand developed by the University of Tokyo's Ishikawa Oku Lab is reportedly so adept at the game rock, paper, scissors that it is unbeatable against a human opponent
4/4
A robot hand developed by the University of Tokyo's Ishikawa Oku Lab is reportedly so adept at the game rock, paper, scissors that it is unbeatable against a human opponent
View gallery - 4 images

A robot hand developed by the University of Tokyo's Ishikawa Oku Lab is reportedly so adept at the game rock, paper, scissors that it is unbeatable against a human opponent.

Strictly speaking that should be robot hand and eye, since, as you might have guessed, it uses a high speed vision system that identifies, within a millisecond, the move being made by the human player from the shape of their hand.

Combined with the clearly very highly responsive actuators in the robot hand and wrist, the robot is able to throw its shape before the human player has finished their move.

Interestingly, this has widely been interpreted as "cheating" on the robot's part by the technology media, but is it, really? No one says human players have to play blindfolded, so really the robot has the same information available to it as does the human. That it can see, process information and respond in the time it takes a human to think of and make a move is hardly its fault.

Ironically, the Ishikawa Oku Lab researchers call this an experiment into "human-machine cooperation." It feels more like a demonstration of machine-human kick-assery, and I, for one, welcome our new rock, paper, scissors overlords.

See video of the robot winning again and again below.

Source: Ishikawa Oku Lab, via IEEE Spectrum

Janken (rock-paper-scissors) Robot with 100% winning rate

View gallery - 4 images
3 comments
3 comments
Mark Dixon
Ah, but what about "rock, paper, scissors, lizard, spock?"
Non-techie Talk
Yes, I think it is cheating.
The nuance of the game is about guessing the most probable best option, or using a gambit (like avalanche, paper roses, scissor sandwich, etc.) to impose on the opponent's psyche, or recognizing the gambit being thrown at you and using the right counter.
The essence of "game" is that either side has the potential of winning.
The computer is not guessing or throwing gambits, it knows the option its component is going to throw, and is selecting the winning option over it. There's no "gamesmanship" here because the computer can't lose.
Consider how computers can't win every chess match, because the computer cannot anticipate what the opponent's next move "will" be (even though it can calculate every available move).
And, remember, having access to information doesn't legitimize leveraging that information. Consider the concept of "insider trading" - someone can have access to information, but it's against the law to use it to profit as it creates an unfair advantage over others in the "stock game."
As another comparative, look at information that is deemed "inadmissible" in a court proceeding. Yes, there is information, and access thereto, but if it's not accessed properly, it is not appropriate for the process.
This very neatly is all summarized by the problem with objective rationality - just because you "can", doesn't mean you "should."
http://nontechietalk.blogspot.com
Mr T
Of course it's cheating Non-techie is exactly right. It is a rather pointless exercise to be honest, can't these guys come up with something better to do with their time, like developing a working bionic arm for amputees?