How Computers Can THINK FOR THEMSELVES

Techquickie ·Techquickie ·2019-05-06 · 908 words · ~4 min read
Floatplane YouTube

Transcript

JSON SRT VTT 76
0:00 so at some point after robots take all of our jobs they'll eventually decide
0:04 that the world just really doesn't need humans that badly and will do away with
0:08 us all together but then haven't you ever wondered how
0:13 this incredibly rosy vision of the future will come to pass
0:16 well i can tell you that it'll probably involve neural networks now we tend to
0:21 think of artificial intelligence as the
0:24 capability of a computer to make sensible decisions on its own kind of
0:29 like a human and as it turns out it's more human-like than you actually might
0:34 think computer scientists and programmers have modeled the way that
0:39 computers arrive at their decisions after our own brains at least to a point
0:45 so inside your nervous system your neurons are arranged in such a way that
0:51 after information is taken in one group of neurons will pass that information to
0:56 the next group in a way that depends on what signals the previous group was
1:01 sending okay so for example then if neurons in
1:05 your muscles sense that they're working really hard because you're going out for
1:09 a run those neurons send out signals through a
1:13 specific pathway that ultimately results in your brain telling your diaphragm to
1:19 breathe harder rather than you know send a message to your hair to grow
1:24 faster and although the way that our neurons
1:27 work is immensely more complicated than
1:30 an artificial neural network made up of ones and zeros the neural networks used
1:35 in machine learning are at least loosely based on the workings of your brain
1:41 instead of neurons made up of dna and cytoplasm though a software neural
1:46 network uses processing nodes stacked
1:49 into many different layers so a node in
1:52 the first layer which initially receives the data
1:55 weights all of its incoming connections by multiplying each piece of data by a
2:01 certain factor depending on the connection it then adds these numbers
2:05 together and if it's above a certain threshold the node sends it along to a
2:09 node in the next layer and then so on and so forth
2:13 once the data gets to the end of the neural network the network has
2:17 effectively made a decision based on
2:20 weighting this combined and then recombined data across many of these
2:25 layers now all of this probably sounds quite
2:29 abstract but basically neural networks use thousands or
2:34 millions of these nodes to turn those numbers into something useful
2:39 of course even the most sophisticated neural network isn't going to know right
2:44 off the bat how to differentiate a picture of a burrito from a body pillow
2:49 just like a baby doesn't pop out of the womb already knowing how to speak the
2:53 king's english so instead developers train a neural
2:57 network by feeding it lots and lots of both relevant and irrelevant inputs so
3:04 for something like a self-driving car it might be a ton of images including some
3:09 of pickup trucks as the data gets fed into the network
3:13 developers monitor its behavior and then adjust the weights on the processing
3:18 nodes until the desired output is reached and the system can eventually
3:23 quickly differentiate a truck from a bicycle or baby stroller or stationary
3:29 object or what have you the whole idea is very general purpose
3:34 so it can actually be adapted to lots of different situations
3:38 other than the more well-known examples that we've already discussed like image
3:41 recognition and autonomous vehicles neural networks have been used for
3:45 everything from filtering spam out of email to training computers to play
3:50 video games nearly as well as professional level human players
3:55 of course training a computer to think is more complicated than getting it to
3:59 run a bunch of pre-defined instructions and as such training a neural network
4:04 can be quite time consuming and depending on the application requires
4:08 some pretty powerful hardware in fact the computers in current self-driving
4:13 cars are not only more powerful than your standard desktop pc they are
4:18 notorious battery hogs but as processors
4:21 become more efficient and our training methods become less tedious we fully
4:26 expect that one day our cars and spam
4:29 filters might be even smarter than we are so even if they do decide to turn on
4:34 us take over the world and rule in tyranny at least we won't have to deal
4:39 with scam emails from nigerian princes or terrible drivers anymore
4:43 private internet access supports a variety of vpn protocols and types of
4:47 encryption and authentication to allow you to dial in the exact level of
4:51 privacy protection that you need they've got apps for Windows macOS Android iOS
4:56 Linux and google chrome with support for other platforms coming you can connect
5:00 up to five devices at the same time with a single account and their apps include
5:04 features like dns leak protection ipv6 leak protection and an internet kill
5:09 switch that will block all traffic if the vpn becomes disconnected
5:13 unexpectedly so check it out today at the link in the video description
5:18 so thanks for watching guys like dislike check out our other videos leave a
5:21 comment if you have a suggestion for a future fast as possible episode and
5:25 don't forget to subscribe to techwiki and ring the bell you look really
5:30 closely make sure it's wrong ringed
5:33 wrong right