Microsoft Taught AI How To Beat 'Ms. Pacman'

Status
Not open for further replies.

dstarr3

Distinguished
I suppose it is kind of scary to think that we already have AI-controlled cars when the most sophisticated AI we have is just now getting a grip on arcade and board games. We might be jumping the gun slightly on the whole AI-controlled car thing.
 


Cars are different due to the radar sensors, the same ones used to tell you when you are getting clos to hitting an object, and the cameras which can interpret and analyze the data in the image in real time.
 

dstarr3

Distinguished
It's not so much about the sensing technology as much as it is the decision-making technology. It reminds of the interesting philosophical dilemma that an autonomous car could face: An accident occurs in front of the the car, and not avoiding the accident would surely kill the passengers in the car. However, the only way to avoid the accident is to drive onto the sidewalk and kill pedestrians. So who does the computer decide to kill and how?

If computers are just now figuring out Pacman, I don't know if they're ready for problems as difficult as that one.
 

sh4dow83

Distinguished
Jul 4, 2011
59
0
18,630


Oh please - as if humans have the capability of figuring out such a problem. You said it yourself - it's a dilemma.

And that's just the ideal case. That somebody will sit there, think "That's quite a problem" and ponder it.

But there's no time!!
So what happens instead?
Human instinct to protect oneself? Mowing down a bunch of kids because even someone who is terminally ill can't think clearly in that split second?
Maybe hesitation - possibly ending up in a whole big family most of whom still have their lives ahead of them in that car getting killed instead of "just" one very old person?

Which I find raises the question - humans are so flawed that I wonder whether machines could on average ever make decisions that are worse. Because even if it would make the decision purely based on number of casualties - most of the time, that's probably a good guess.

Of course... knowing our screwed up world, I suspect that somebody would put in a whitelist for special rich people with special implants sooner or later...
 
If a kid jumps in front of your car from 20 feet away while you are going 40 miles per hour (56 feet per second) the 2 objects would collide in about 0.35 seconds.

For comparison a blink is 0.3 to 0.4 seconds.

An unlucky/inattentive human driver may not have time to react at all and collide at full speed.

A lucky/experienced human driver may break hard and swerve out of the way into a randomly chosen lane, left or right, without caring about any cars on either side of him knowing the cosmetic damage to a car can be repaired but a person is much harder to fix.
This driver may not hit the kid nor collide with any other vehicle meaning all obstacles are avoided and no damage was done to anything / person, of course most drivers would be visibly shaken by this event.
The issue is that not everyone has the experience, reaction time and reliability to do this every single time.

A computer driving a car would know the car's exact limits rain or shine along with the current traction on each wheel.
I remember the 2016 GMC Sierra Denali commercial "1000 times a second".
Thus allowing the car to either swerve around the kid completely or maneuver the car into a nearby open lane, being able to monitor each lane and even a complete 360° around the vehicle.

A group of AI powered vehicles could even communicate with each other to help one of them avoid an obstacle.

Say car A and car B are driving side by side on a 2 lane road, 2 one direction a medium and 2 in the opposite direction.
A kid jumps into the path of car A.
Car A asks car B if it is possible to slow down so I can immediately take your position to avoid this obstacle.
Car B says sure and instantly applies its brakes.
Car A then swerves into car B's lane avoiding the kid.

AI may would make for faster trips as well.

Think of a red light.
When the light turns green all the cars don't go at once, they go in sequence.
Car 1 releases his brake and accelerates, car 2 sees the brake lights on car 1 have disappeared and in 0.3 seconds car 2 releases his brakes and pushes the accelerator and this process goes on and on till it gets to your car.

With AI as soon as the light turns green all the AI vehicles could accelerate at a steady rate all at the same time.
Technically the acceleration rate does not have to be steady it just has to be the same for each car so none of them "move" in relation to each other, always 15 feet in front and 15 feet behind for all cars meaning you could have 10 Tesla Model S P100D accelerating a full speed in a line and as long as each one of them maintains the same acceleration everything is fine, although I'm sure some regulatory company will institute a framework, maximum acceleration, maximum speed, minimum space between vehicles and of course all different variables for when it's raining, we still have to obey the laws of physics.
 

urbanj

Honorable
Dec 27, 2012
121
0
10,710



Thing you forgot about in a world where ALL cars are AI, is that there would be NO RED LIGHT :p

...we might also consider not having windows at all in vehicles at that time either, as the sight of our AI driven cars zipping past one another within fractions of an inch at high speeds would like cause most people to change their pants upon exiting the vehicle :lol:
 


That is true, if all cars were AI then we wouldn't need red lights, but unless we pass a law making it illegal to drive there will still be people driving who will need a red light and with our cars not recognizing them as fellow AI would most likely slow down around them while we give them the American Salute.
 
A far more likely scenario, at least in the next couple decades, are humans sharing the job of driving with AI rather than being totally replaced by them.

That above line does feel like a speech they would give before presenting the prototype of Skynet
 

dstarr3

Distinguished


Well, the biggest difficulty is that autonomous cars are going to need to share the road with "analog" cars, for a very long time. The whole thing would be easy if suddenly every car ever was autonomous, then we'd just never have accidents or traffic because every car would be connected and aware of everything everywhere. But for many, many decades to come, there will be at least some human-controlled cars, and with that comes the chaos of machines trying to navigate around unpredictable humans.
 

LORD_ORION

Distinguished
Sep 12, 2007
814
0
18,980
I can't help but notice they used Atari 2600 Ms Pacman. The AI is very simple and always makes the same choices. All they've basically done is find a route to optimal pattern recognition.

This was the problem with the original arcade Pac Man... and why arcade Ms Pacman has a more erratic AI (to eat more quarters, instead of let smart players play for an hour at a time), in that there is a randomizer in its ghost decisions that make a pattern nearly indecipherable, because the ghosts won't always behave the exact same way given any game state.

Having an AI that could navigate Ms Pacman arcade would be immeasurably more impressive.
 

blackbit75

Distinguished
Oct 10, 2010
49
0
18,530
What if AI went claimming their rights? a body in a robot to live in, or their own country to live, asking not being turn them off because they don't want to sleep.
 

McWhiskey

Reputable
Sep 23, 2015
41
0
4,530
The discussion related to AI in cars kind of missed the point of the OP. All of the rebuttals were describing situations outside of the intent of the situation. In fact, they were just reiterating the perceived benefits of AI in ideal scenarios. The question is, what does the AI choose when the situation is not ideal?

With a human driver, life keeps moving forward. The driver panics and locks it up, or panics and swerves, or any number of other outcomes. Either way it just happens and it is spontaneous. The outcome is regrettable and sometimes understandable. But it is a thing that "just happened" that way.

But for an AI the choice is not spontaneous. Values have to be placed on everything. The article even states this. Avoiding a ghost trumps eating a pellet. How does an AI choose the values? Is it predetermined by programmers? Who could possibly be expected to code morality? Yet it still needs to be done.

My take on the OP is that we are running head first as fast as we can into a complex unpredictable world with self driving cars but we are still just chipping away at solving rudimentary yet tough decisions.
 


If the AI never went to sleep then how would he dream?
All intelligent beings dream :)

 
About DRIVING: self-driving cars will NOT be available en masse until they are better than humans. Cut and dry. In fact, there are already cars that record speed, braking time etc which are then analyzed in testing, and after REAL accidents.

We are also continuing to get ASSISTED driving such as auto-braking if we get too close to an object (that exists today).

As for the actual software, to over-simplify the cameras send a video to the computer which in turn breaks down the image into its parts; this research was based on analyzing how young CHILDREN think; in fact it turns out to be the first stage in human-like artificial thinking. The next step is contextual awareness such as "boy on bike moving down hill" which could trigger an ALARM (though again, anti-braking could still kick in).

So basically we need to know where the ROAD is at all times, where we should be on it, when we should slow down or stop based on simple cues as well as complex, contextual-based analysis. Examples of STOPPING include evasion of obstacles that can't be avoided by slowing or swerving safely, unsafe driving conditions, or errors detected in the cars hardware.

Finally, even after cars come to market en masse the data will continue to be analyzed to improve self or assisted driving.
 
BTW, the most IMPORTANT thing in the future in terms of economics will be evolutionary based algorithms that solve problems such as engineering and medicine. For example, you could tell the computer to solve a persons cancer. The PROCESS for this is to define the SOLUTION in enough detail. Huh?

The computer would know the basic parameters for what a HEALTHY cell is. You would then give it a digital version of a persons unhealthy cell (a 3d picture which probably isn't possible yet). The computer could simply give it various combinations of molecules then record how long the cell lives.

The EVOLUTIONARY part of the algorithm works thus.. if a straight saline solution makes the cell live say 60 days, then tray saline with something else. Record time. Then give PRIORITY to variations of a solution that make the cell last longer, and LESS time to solutions that do not. (The "answer" would need to be suggested based on healthy cell lifespan)
 

numach

Reputable
Sep 30, 2015
5
0
4,510
Am I the only person wondering why they chose the Atari VCS version of Ms. Pac-man instead of the actual game?
 
Status
Not open for further replies.