DeepMind Technologies, a company that develops machines that try to learn as humans do, has published a paper showing the machine has learned to play games on the old Atari 2600. It may not sound like a big deal, except the machine learned only by evaluating the pixel on the screen and had no other inputs. No rules, no game-specific programming, nothing but what the screen outputs. The amazing thing is that it adapted to the visual cues to learn the games just as a human being would do.
Space Invaders on the Atari 2600
The games weren't terribly difficult: BeamRider, Pong Sports, Breakout, Space Invaders, and Seaquest. More complex games on the console, like Q*Bert, still cause it to stumble. But it's ability to deduce rules solely from screen output is remarkable and unnerving at the same time, and shows promise.
Google recently acquired the company to help it parse data from images, and all large tech companies including IBM, Microsoft, Apple, and others use deep learning to help their systems perform better whether it's financial analysis or speech recognition.
Professor in this case I think I get more scared than gladly surprised. It gets to that point in class in which were talking about AI and how we want and need computers but don't want them to get to the point in which they have their own "conscience". This kind of new technology keeps taking us one step closer and closer to that level. Now looking at the bright side, I can imagine the happiness that the scientists developing this intelligence must feel to be able to produce the first AI that can do something out from other outputs and not input, very impressive.
ReplyDeleteThis raises deep concern for me. Maybe it is just me, but I do not want any type of electronic system to learn and set its own rules and parameters. Computer systems are great at doing what we program them to do. If a fault or error occurs we can look at the programing and fix any error. I think it should stay that way. A deep learning system, however, would rapidly create its own programing at an alarming rate. Slowly, all the Sci-Fi movies are becoming reality; specifically IRobot and Terminator (as you stated in class).
ReplyDeleteI am mildly concerned. I for one believe that this technology is not only coming it is here. I see these developers, working more towards refining and tuning. There are some benefits. Wars may be fought without human casualties. We would be able to have some work done that would be dangerous to humans, especially where making a decision is needed. (example fire fighter) Instead of putting a human in the way, it would be a benefit. The way I see it, there are military and financial benefits in the future. Just a thought.
ReplyDeleteI am mildly concerned. I for one believe that this technology is not only coming it is here. I see these developers, working more towards refining and tuning. There are some benefits. Wars may be fought without human casualties. We would be able to have some work done that would be dangerous to humans, especially where making a decision is needed. (example fire fighter) Instead of putting a human in the way, it would be a benefit. The way I see it, there are military and financial benefits in the future. Just a thought.
ReplyDeleteI am on board with the other student comments on this one. A computer system that can actually learn from simply watching screen pixels is super scary. I guess I can see some practical applications as Ainsley pointed out, but I hardly want to be replaced by a computer. As it is typically I find that the lack of human interaction is a little troubling even with just simple customer service transactions. What is next in artificial intelligence? And is it worth what humans could risk to have this type of technology?
ReplyDeleteNichole,
DeleteIs it really learning, or is it recording and using probability algorithms to predict your next move. I think learning is different. Learning is having the ability to make independent judgement.
I think the fear of being replaced is very far, far, far away if at all possible. I never say never though. I love your question "is it worth what humans could risk to have this type of technology"?
I have been watching a show called Almost Human. Check it out. I am curious to get your opinions on it. I think it is similar to our class discussions.
The development of Artificial Intelligence may have already been predicted to be achieved now and be developed more in the near future. With the technologies, it's only a matter of time where Artificial Intelligence will replace the jobs that are performed by human beings. It's just a scary thought that the possibility of Artificial Intelligence taking over. What will happen to us? Although, it's the "future" it just made me think that even the entry level positions can be affected for example, working at a fast food chain. What if they develop something that can flip burgers and cooking it well by detecting the temperature of the burger, etc. I believe that if the development of Artificial Intelligence becomes more advanced, it can affect us and our jobs that may lead to high unemployment rate and poverty. So, I hope they're considering how it'll affect other people when developing something new.
ReplyDeleteI find this be a little frightening and exciting all in one. Yes, its great that our technology is consistently evolving, but on the other hand human instincts aren't something a computer could have programmed. I'm not sure I could trust a computer or robot for that matter to do what it is programmed to do, especially in a War. Computers malfunction all the time! It is a double edge sword technology helps to simplify things, but it also takes away from human interaction.
ReplyDeleteThis comment has been removed by the author.
ReplyDeleteI think its amazing that technology has made it this far and these types of breakthroughs will help human beings live better lives. I just hope we don't get to dependent on it and allow it to cause a negative impact on the human race. If we have machines take over most of the workforce, leaving only a short percentage of the human population to control them, what happens to the vast majority. Maybe we can all share the responsibility?
ReplyDeleteAlso, what if something goes wrong with a machine and it starts attacking a human being or better yet a child. There better be a very good "fail safe."
I honestly do not believe that this is a good idea. This might sound a little funny but we are coming closer and closer to a robotic take over. I do not understand why we want "robots" to learn. The way society is right know I think is fine. Why do we want to get carried away; seems to me that the human race is relying more and more with electronics. The movies terminator might be more realistic then science fiction. If robots begin to learn on there own sooner or later they will realize that the human race is obsolete compared to them.
ReplyDeleteOn the other hand it could be good just because it can make our lives easier. For example if the robot is "learning" we can teach it to go to the store and buy groceries or medication from a pharmacy. This is the begging of either great technology or the beginning or to our extinction.
This type of technology is leading us into another era. The robotic intelligence era. If we keep finding ways to incorporate robots into the work force then we are going to start losing some of the old school way of doing things. No matter how we look at it, machines can only be programed by commands they will never be able to take on a thought process on their own. Even though this technology seems to real we are the creators. It is taking us to bigger advantages for the future. I look forward to what machines will be able to do for us in the future.
ReplyDelete