Artificial intelligence, video games and the mysteries of the mind | Raia Hadsell | TEDxExeterSalon

  Рет қаралды 19,429

TEDx Talks

TEDx Talks

Күн бұрын

Пікірлер: 53
@anubisparadox7105
@anubisparadox7105 7 жыл бұрын
The overwhelming feel of Success the blinding moment and the realization of a mistake
@Djolewatchtastife
@Djolewatchtastife 7 жыл бұрын
By teaching AI how to learn and understand, we would surely discover more effective ways of learning and understanding concepts as humans.
@DriftyG
@DriftyG 7 жыл бұрын
The title alone had me like, yup!
@anubisparadox7105
@anubisparadox7105 7 жыл бұрын
And to enjoy losing because it’s apart of the process of learning and information is the true goal via experiences we definitely have in the honest community something we call a beautiful or good lost
@hsk8787
@hsk8787 6 жыл бұрын
They may solve some of our problems but they may be the end of us. If they have the ability to continually learn and develop themselves, it will surely be only matter of time before they overtake us and determine they should be in control.
@m_sedziwoj
@m_sedziwoj 7 жыл бұрын
We forget how to play games too, but much slower.
@jas0ndrinkbeer8
@jas0ndrinkbeer8 7 жыл бұрын
Technology is so amazing that it kinda scares me...
@sattyre6892
@sattyre6892 2 жыл бұрын
I really hope you guys and gals are putting equal thought into containing and or limiting AI along the way. There is a lot of science fiction illustrating both the positive and negative ramifications of AI, and we, meaning Humanity, only get one chance at this. Lets be careful in how we create, contain and manage the developement of AI. Because if the negative comes to pass, sorry is going to be meaningless.
@lordphu
@lordphu 6 жыл бұрын
She talks like Leonard Hofstadter's Mom!
@sourabhkhandelwal689
@sourabhkhandelwal689 5 жыл бұрын
and also has similar looks.
@rogerguo8352
@rogerguo8352 3 жыл бұрын
exactly!!!
@anubisparadox7105
@anubisparadox7105 7 жыл бұрын
Did you teach it to enjoy it self? Whenever we are angry we try to be the best at war games, but wen we are calm we play mind games? For example....
@anubisparadox7105
@anubisparadox7105 7 жыл бұрын
Just sit back and look at the waves come and go....
@hereb4theend
@hereb4theend 5 жыл бұрын
Lisp + prolog + Haskell === Julia
@monstersince
@monstersince 6 жыл бұрын
online games use Ai to constantly assess players and can independantly 404 your gaming account
@skydivekrazy76
@skydivekrazy76 6 жыл бұрын
The scarry thing about AI is that it's being driven by big business. And big business destroys everything in its path to achieving its own selfish goals. Primarily because most big business are controlled by social paths and psychopaths... Look it up, lots of research to prove this. And AI is being modeled to follow this example. Anybody else fear this reality?
@ivang5009
@ivang5009 5 жыл бұрын
Yeah, one day you will be playing chess vs the computer and it will win you, maybe you also loose at pong
@anubisparadox7105
@anubisparadox7105 7 жыл бұрын
A computer will always lose wen it comes to enjoying the moment because of overwhelming mass of information
@anubisparadox7105
@anubisparadox7105 7 жыл бұрын
Fun
@anubisparadox7105
@anubisparadox7105 7 жыл бұрын
Sometimes we lose on purpose
@harpsingh7350
@harpsingh7350 6 жыл бұрын
3 people don't like A. I.
@antoniobento2105
@antoniobento2105 2 жыл бұрын
This is the role model I would want my daughters to have, not Cardi B, not Meryl Streep.
@gotreachyet
@gotreachyet 3 ай бұрын
2024 here it did not take decades
@artemkovera5500
@artemkovera5500 6 жыл бұрын
ALL those effects could have been achieved by using genetic algorithms.
@sourabhkhandelwal689
@sourabhkhandelwal689 5 жыл бұрын
Neuroevolution is used extensively to train AI to plays games.
@foxxxof
@foxxxof 6 жыл бұрын
I think expecting one trained AI to do multiple things is the wrong approach. I always thought what would be better would be to create a vast number of trained AI all on varying tasks, and then have another AI that is good at guessing which ones would solve a new task well, or that could take examples of other AI to create a new one that is good at a new task. Much like a bunch of specialists and a good manager. Sure you might need a ton of computers to do that, maybe even a whole server room or warehouse of computers all dedicated to it, but how is that different than our first computers?
@siddharthnilawar
@siddharthnilawar 4 жыл бұрын
Generalization is always better... And the goal of DeepMind is to create an AI that can learn by itself and think like humans do... That would mean centuries of progress in few days...
@andres6868
@andres6868 7 жыл бұрын
ai will solve inequality? on the contrary, it will increase inequality to never seen before levels, as most of the wealth created will go to the few owners of the machines.
@sofia.eris.bauhaus
@sofia.eris.bauhaus 7 жыл бұрын
do you think this would be a smart thing to do?
@robertkelly9772
@robertkelly9772 7 жыл бұрын
No, it's not the smart thing to do but humans will do it any how. It's not because humans aren't smart it's because they opt not to do the smart thing. Humans are a perverse species.
@sofia.eris.bauhaus
@sofia.eris.bauhaus 7 жыл бұрын
well, one of the reasons for ai is to supersede the shortcomings of the human mind.. which species would do you prefer?
@andres6868
@andres6868 7 жыл бұрын
as a human, why I want to be superseded?
@sofia.eris.bauhaus
@sofia.eris.bauhaus 7 жыл бұрын
Andres Karel: you probably don't. but you might want more intelligence on earth.
@sofia.eris.bauhaus
@sofia.eris.bauhaus 7 жыл бұрын
tunneling in breakout is not a superhuman ability. just saying..
@foxxxof
@foxxxof 6 жыл бұрын
I figured that out in less than 3 hours too... when I was maybe 5 or 6. Just saying... These folks must be terrible at video games.
@robosergTV
@robosergTV 6 жыл бұрын
So it took you to do 5 years what the system learns in couple of days. Also superhuman means it gets more points then a human. I doubt you can do better even now. Soon you will be obsolete for a real job anyway.
@robosergTV
@robosergTV 6 жыл бұрын
they get better score then an average human, so hard to understand that?
@Fiscus128
@Fiscus128 6 жыл бұрын
She hopes that AI solves or will contribute to solving inequality. I think there is a huge gap between what she hopes or likes to see and how reality looks like or will look like. The ending of this presentation unfortunately weakens the generally interest8ng first 95 %.
@robosergTV
@robosergTV 6 жыл бұрын
How is making goods and services cheaper not reduce the inequality? Average Joe now has better medical help then kings some 100 years ago. People like you sadden me.
@ravenxrgaming4672
@ravenxrgaming4672 3 жыл бұрын
It's because you might not know how an AI works.
@Fiscus128
@Fiscus128 3 жыл бұрын
@@ravenxrgaming4672 Enlighten me
@ravenxrgaming4672
@ravenxrgaming4672 3 жыл бұрын
@@Fiscus128 AI is like an imitation of our own mind but we can leverage its power to a maximum for maximum yield and gets into solution that our humans can't even figure it out. Eg:- in this pandemic many companies like nvidia are using AI for learning and predicting the mutation of the covid 19 virus. It helps for finding specific vaccines against it. So yeah it good. But there is a universal truth exploiting any resource or technology in large scale is always bad. It can be applied to anything. Even the food we eating you know what I am saying.
@Fiscus128
@Fiscus128 3 жыл бұрын
@@ravenxrgaming4672 "AI is like an imitation of our own mind " It sounds nice, but the reality is that humans actually have a very poor idea how the brain actually works, let alone that AI is an imitation of it. The moment AGI really exists is also the moment where we - as a species - may become really worried about our existence.
@bjpcorp
@bjpcorp 7 жыл бұрын
She lost me at "AI will EASE inequality."
@anubisparadox7105
@anubisparadox7105 7 жыл бұрын
Lol
@donfields1234
@donfields1234 5 жыл бұрын
Really...THE AGENT...did she see the matrix??? Lol
@ONDANOTA
@ONDANOTA 6 жыл бұрын
yet another rip-off of a Demis Hassabis talk
What would a conscious AI look like? | Heather Berlin | TEDxKC
18:05
PRANK😂 rate Mark’s kick 1-10 🤕
00:14
Diana Belitskay
Рет қаралды 9 МЛН
How Strong is Tin Foil? 💪
00:25
Brianna
Рет қаралды 68 МЛН
ТВОИ РОДИТЕЛИ И ЧЕЛОВЕК ПАУК 😂#shorts
00:59
BATEK_OFFICIAL
Рет қаралды 2,8 МЛН
这是自救的好办法 #路飞#海贼王
00:43
路飞与唐舞桐
Рет қаралды 127 МЛН
AI & The Future of Work | Volker Hirsch | TEDxManchester
18:21
TEDx Talks
Рет қаралды 541 М.
The Deep End of Deep Learning | Hugo Larochelle | TEDxBoston
18:00
Putting Yourself Out There | Lori Granito | TEDxLingnanUniversity
50:01
PRANK😂 rate Mark’s kick 1-10 🤕
00:14
Diana Belitskay
Рет қаралды 9 МЛН