Read more: andauth.co/AIvsML | The terms artificial intelligence and machine learning are often used interchangeably these days, but there are some important differences. Let me explain.
Download the Android Authority App: https://andauth.co/aa-app
Subscribe to our YouTube channel: https://andauth.co/subscribe
Music: Last Summer by Ikson https://soundcloud.com/ikson
Creative Commons — Attribution-ShareAlike 3.0 Unported
Stay connected to Android Authority:
Follow the Team:
Josh Vergara: https://twitter.com/jvtechtea
Joe Hindy: https://twitter.com/ThatJoeHindy
Lanh Nguyen: https://twitter.com/LanhNguyenFilms
Jayce Broda: https://twitter.com/jaycebroda
Gary Sims: https://twitter.com/garysims
Kris Carlon: https://twitter.com/kriscarlon
Nirave Gondhia: https://twitter.com/nirave
David Imel: https://twitter.com/durvidimel
Bailey Stein: https://twitter.com/baileystein1
The neural network doesnt "start off all zeros". I'm not sure if this was just a slip of the tongue but this is absolutely not what happens, the neural network would not train properly if it was initialized to all zeros.
Machine learning is a subset of AI. They are not "separate" things. As someone who studied and does work in this area, your explanation is not quite accurate. To people who watch this and take it for Gospel. Do your research.
Video is great. Thank you. But please please please don't put any music into background in your videos. I don't know what is the reasoning on adding background music. If you think we need to be entertained when we listen to these subjects, you are wrong. Because we are not bored. Would you play a stupid background music during a TedEx talk? You are degrading the quality of your videos. Lowering it to a "top ten" quality.
Are you ok Gaz ? looking a bit burned on the scone mate , reminds me of a time i was having a beer on the beach and fell asleep in the sun , i looked like a beetroot for about two weeks .
hope thats not painful for you .
gary FOR FKN PRESIDENT! its been a while since i posted this. but i posted other fanboy shit in the meanwhile. i just LOVE your stuff on this channel. the guys are hopefully proud. greetings from germania.
Instead of asking “oh please explain it to me”, Gary says “please let me explain!!!” with an insane amount of detailed graphics so it looks just that simple.. That’s ridiculously beautiful.. BUT some stupid people who disliked it, don’t appreciate a precious information that is given for FREE!!! I don’t understand!
garry im confused
why do phones have up to 20 gpu cores and pcs only 1.
and why is my gpu always only used up to 30% but my cpu up 70% thats pretty bad balancing.
So why so many gpu cores
could you do a garry explains vidio if you wont answer this comment .
Hey Gary, just clicked on the video because I wanted to know why there is my public library in the background =) (Hoped a little bit that you're in town right now ^^) Greets from Germany - Stuttgart, Jo
I would say that we don't have any examples of AI yet. We have some sophisticated systems which are based mainly on machine learning and are very good at specific tasks. However general AI is not here yet. Having said that, it does depend on how you define these things exactly. If someone said that today we do have narrow tasked weak AI, I guess I wouldn't argue.
Well that was the point of the video, so I guess I failed there! Voice assistants are just lookup systems. Mainly based on machine learning to recognize the voice input, then recognize the language and then look up a suitable reply. No "intelligence".
I use a different background for every video, so technically there is no "old" background. Unless you are talking about the black & white cityscape that I have on the wall of my studio? But I haven't used that for ages, we are in the green screen age now!
to make computers aware of thare users in much the same way a human becomes aware. artificially aware of things sounds and people around sounds great. the fact we are allready using desktop pcs with more tflops of compute performance than whats depicted as what data from star trek has yet recognition of things people and sounds around are still not even compairable to an ape never mind a human. so whats called a neural net is to make a chip like a gpu that has over 4000 stream processors and 20-25 tflops of compute power and rearrange the physical layout of the parallel processing cores so its no longer in a grid pattern to make something that works better not faster becouse raw computer power is not cutting it with even basic recognition of things sounds and people. just think intel is going to make cpus this year that have more horsepower than data from star trek and gpu makers have been making parallel processors with that kind of power since 2012 and none of that is cutting it.
2:41 No, not exactly you're assuming here. Yes, AlphaGo does use a tree search algorithm to find its moves from "learned" information by machine learning, but it can call upon its artificial neural network. Same as to human players, the new rules would need to be explained the only difference is the board size. The human would still lose probably the first game.
Otherwise, your explanation would be like saying an Autonomous Driving AI would crash upon its first encounter of a e.g. 10 lane highway assuming the prior was an 8 lane highway.
So you're taking a semantics argument?!
I could argue that there isn't an AI of any kind yet since no one can determine exactly what intelligence is exactly and certainly not what it means to be self aware, etc and I'm certainly not claiming AlphaGo is an AGI as such as it is currently. However, DeepMind and AlphaGo are taking the elements of learning and applying them artificially. AlphaGo unarguably doesn't solve (play) problems on Brute Force algorithms alone, and if the current version of AlphaGo can learn by itself it unarguably is a step forward for AI and the beginnings of AGI all be at a weak AGI.
OK, I give up. The AlphaGo Zero paper published in nature and the blog articles on deepmind.com state many times over that this is a reinforcement learning system.
There is even a link to a Wikipedia page about RL, which start like this, "reinforcement learning (RL) is an area of machine learning..." I have dozens of Machine Learning example program here on my PC that use RL and unsupervised learning. They are both well known machine learning techniques. If you think that AlphaGo is a General AI then I can't help you.
Gary .. it doesn't .. that's the whole point. Your information is incorrect. AlphaGo isn't a Checkers, Chess or "garbage in garbage out" specific 'AI' .. instead it's a General AI an early one to what we project AGI / ASI's to become.
You also need need to understand the game Go .. there are a few go ones published.
I recall our last argument over cryptocurrency .. and now China is one of the first counties to ban them. I also remember similar arguments over Chinese Android phones and now AT&T and Verizon are starting with Huawei. Nevermind, the 5G concerns and shakedowns.
Gary .. I'm not some dumb guy trying to argue.
D Jaquith I am Gary, but that aside, you are still missing the fundamental idea here. Even though AlphaGo doesn't need human intervention, it still has to have a training phase where it plays against itself and learns how to play the game. With that established data set it can then play others. If that data set is used to play on a different board size it will fail. It can be retrained with the new board size, but it is a machine learning system, it doesn't "understand" Go and it can't adapt it's "thinking" to a new board size on the fly. It would need to develop that experience from scratch.
Vitale Family Law offers North Carolinians the opportunity to resolve family law challenges in a personalized, effective manner. No one wants to waste time or energy on futile efforts. Working with experienced attorneys inspires confidence and helps move cases toward resolution. Lawyers at the firm have prepared diligently to meet the needs of clients facing some of life’s most difficult transitions, including divorce and child custody disputes. Schedule an initial consultation to discuss your situation with an experienced, well-recognized Raleigh law firm.