{"id":3571,"date":"2023-01-10T00:59:12","date_gmt":"2023-01-09T21:59:12","guid":{"rendered":"http:\/\/www.artificialbrain.xyz\/?p=3571"},"modified":"2023-01-13T01:17:40","modified_gmt":"2023-01-12T22:17:40","slug":"towards-artificial-general-intelligence-oriol-vinyals","status":"publish","type":"post","link":"https:\/\/www.newworldai.com\/towards-artificial-general-intelligence-oriol-vinyals\/","title":{"rendered":"Towards Artificial General Intelligence: Oriol Vinyals"},"content":{"rendered":"<p style=\"text-align: justify;\"><strong><span style=\"font-size: 14pt;\"><em>What is artificial general intelligence and what are researchers doing to achieve this goal today? Oriol Vinyals, from Google DeepMind, walks you through the answers to these questions with some interesting examples.<\/em><\/span><\/strong><\/p>\n<p><iframe loading=\"lazy\" src=\"https:\/\/www.youtube.com\/embed\/vueAUa7OPCg\" width=\"760\" height=\"315\" frameborder=\"0\" allowfullscreen=\"allowfullscreen\"><\/iframe><\/p>\n<p style=\"text-align: justify;\"><span style=\"font-family: helvetica, arial, sans-serif; font-size: 12pt;\">I hope most of you have heard the term of artificial intelligence or artificial general intelligence. The sort of textbook definition would be sort of the study of building agent (as we call them) to solve a a \u00a0large variety of very complex tasks before without needing to be reprogrammed. <\/span><\/p>\n<p style=\"text-align: justify;\"><span style=\"font-family: helvetica, arial, sans-serif; font-size: 12pt;\">These is a very long quest which we have taken on and there&#8217;s still a long way to go for us to sort of see it done. But crucially it&#8217;s already shown a lot of advancements and a lot of progress that we already have seen our day today.<\/span><\/p>\n<p style=\"text-align: justify;\"><span style=\"font-family: helvetica, arial, sans-serif; font-size: 12pt;\">But this thought.. I actually want\u00a0<\/span><span style=\"font-family: helvetica, arial, sans-serif; font-size: 12pt;\">to explain how we the researchers are actually trying to achieve this goal so not only like define it but actually how are we tackling nowadays building agents. <\/span><\/p>\n<p style=\"text-align: justify;\"><span style=\"font-family: helvetica, arial, sans-serif; font-size: 12pt;\">Imagine for a second that I asked you to be sort of a bit of an engineer perhaps and do something that to us looks extremely trivial we do this\u00a0<\/span><span style=\"font-family: helvetica, arial, sans-serif; font-size: 12pt;\">every day.<\/span><\/p>\n<p style=\"text-align: center;\"><strong>&#8220;Make a robot open a door&#8221;<\/strong><\/p>\n<p style=\"text-align: justify;\"><span style=\"font-family: helvetica, arial, sans-serif; font-size: 12pt;\">There&#8217;s a robot and I ask you okay you build an algorithm a\u00a0<\/span><span style=\"font-family: helvetica, arial, sans-serif; font-size: 12pt;\">sequence of instructions to make this robot open the door. It is very\u00a0<\/span><span style=\"font-family: helvetica, arial, sans-serif; font-size: 12pt;\">common to sort of break out this problem into several pieces and what you could do is sort of say do that right so to open up, find the door, find where the handle of the door, grab it, push it\u00a0<\/span><span style=\"font-family: helvetica, arial, sans-serif; font-size: 12pt;\">down and open it.<\/span><\/p>\n<p style=\"text-align: justify;\"><span style=\"font-family: helvetica, arial, sans-serif; font-size: 12pt;\"> I turns out out that if you actually strictly speaking do\u00a0<\/span><span style=\"font-family: helvetica, arial, sans-serif; font-size: 12pt;\">this sort of rule-based and hierarchical way to solve a complex problem breaking\u00a0<\/span><span style=\"font-family: helvetica, arial, sans-serif; font-size: 12pt;\">it down into smaller not so complex problems, you end up with something that\u00a0<\/span><span style=\"font-family: helvetica, arial, sans-serif; font-size: 12pt;\">perhaps is not so optimal.<\/span><\/p>\n<p style=\"text-align: center;\">Find Door<br \/>\nFind Handle<br \/>\nGrab Handle<br \/>\nPush it down<\/p>\n<p><iframe loading=\"lazy\" src=\"https:\/\/www.youtube.com\/embed\/g0TaYhjpOfo\" width=\"760\" height=\"315\" frameborder=\"0\" allowfullscreen=\"allowfullscreen\"><\/iframe><\/p>\n<p style=\"text-align: justify;\"><span style=\"font-family: helvetica, arial, sans-serif; font-size: 12pt;\">If you simply try to program, what&#8217;s going to\u00a0<\/span><span style=\"font-family: helvetica, arial, sans-serif; font-size: 12pt;\">happen? If the door is open, (you are not accounting for these) the result is not\u00a0<\/span><span style=\"font-family: helvetica, arial, sans-serif; font-size: 12pt;\">your fault. The key concept here is that your algorithm\u00a0<\/span><span style=\"font-family: helvetica, arial, sans-serif; font-size: 12pt;\">will not generalize to the complexity of the world. This by the way is from the\u00a0<\/span><span style=\"font-family: helvetica, arial, sans-serif; font-size: 12pt;\">darker competition in 2015. <\/span><\/p>\n<p style=\"text-align: justify;\"><span style=\"font-family: helvetica, arial, sans-serif; font-size: 12pt;\">These programs are not really as simple\u00a0<\/span><span style=\"font-family: helvetica, arial, sans-serif; font-size: 12pt;\">as the four lines I showed you and it really shows that trying to build\u00a0<\/span><span style=\"font-family: helvetica, arial, sans-serif; font-size: 12pt;\">algorithms to these like rule-based chain of rules (sort of approach) is bound to failure\u00a0<\/span><span style=\"font-family: helvetica, arial, sans-serif; font-size: 12pt;\">in mistakes. Of course, these were like the funny ones.<\/span><\/p>\n<p style=\"text-align: justify;\"><span style=\"font-family: helvetica, arial, sans-serif; font-size: 12pt;\">Here is how you could do this perhaps in a slightly different way. I call this the learning approach\u00a0<\/span><span style=\"font-family: helvetica, arial, sans-serif; font-size: 12pt;\">and this is studying a lot knowledge. Take the problem of finding the door. How would you program that? Maybe you have these pixels, you try to find these rectangular objects that look like doors perhaps and as you saw it&#8217;s not gonna channel eyes to any conditions you may have not imagined and then you have to try on error iterate and this really doesn&#8217;t scale because you&#8217;re just now opening doors and how about doing anything else? <\/span><\/p>\n<p style=\"text-align: justify;\"><span style=\"font-family: helvetica, arial, sans-serif; font-size: 12pt;\">The paradigm here is I&#8217;m gonna instead try to get a lot of pictures of doors and I have human labelings where the doors are. What I gonna do is I&#8217;m gonna learn what a lot door looks like instead of programming in terms of rules. It is very very important.<\/span><\/p>\n<p style=\"text-align: justify;\"><span style=\"font-family: helvetica, arial, sans-serif; font-size: 12pt;\">Unfortunately in computer vision, people care about other objects not doors and people generally don&#8217;t take pictures of doors. I am going to show you the example of finding actually cats and dogs which are very popular objects and in the research community. How are we actually gonna do that? <\/span><\/p>\n<p><img decoding=\"async\" loading=\"lazy\" class=\"alignnone wp-image-3621\" src=\"http:\/\/www.artificialbrain.xyz\/wp-content\/uploads\/cat-or-dog-450x253.jpg\" alt=\"\" width=\"802\" height=\"451\" srcset=\"https:\/\/www.newworldai.com\/wp-content\/uploads\/cat-or-dog-450x253.jpg 450w, https:\/\/www.newworldai.com\/wp-content\/uploads\/cat-or-dog-178x100.jpg 178w, https:\/\/www.newworldai.com\/wp-content\/uploads\/cat-or-dog-768x432.jpg 768w, https:\/\/www.newworldai.com\/wp-content\/uploads\/cat-or-dog-1024x576.jpg 1024w, https:\/\/www.newworldai.com\/wp-content\/uploads\/cat-or-dog.jpg 1368w\" sizes=\"(max-width: 802px) 100vw, 802px\" \/><\/p>\n<p style=\"text-align: justify;\"><span style=\"font-family: helvetica, arial, sans-serif; font-size: 12pt;\">The first step to finding that cat whether an image contains (let&#8217;s say a cat or a dog) is to build a big database of pictures that contain either cats or dogs and then literally as humans to label thousands or tens of thousands examples of images like this one which actually is a dog although this one is pretty tricky. I would say so how would you even you know to code these and go through the pictures and so on.<\/span><\/p>\n<p style=\"text-align: justify;\"><span style=\"font-family: helvetica, arial, sans-serif; font-size: 12pt;\">The first piece is to get these databases of labeled examples. Once you have these, the second step is allow for a generic program to be able to learn the characteristics or the patterns on these databases. That&#8217;s precisely what deep learning and deep neural nets are enabling us to do.<\/span><\/p>\n<p style=\"text-align: justify;\"><span style=\"font-family: helvetica, arial, sans-serif; font-size: 12pt;\">here you see a big neural network that actually processes the image that were pixels and activates certain neurons in this architecture that could well be for instance detecting the size of the animal, whether it has big ears and so on..<\/span><\/p>\n<p style=\"text-align: justify;\"><span style=\"font-family: helvetica, arial, sans-serif; font-size: 12pt;\">Crucially, these features that these neural network is activating on this image and not prepared by us. These are fully learned from these database. We have the database. We have this model that is powerful enough to represent all sorts of interesting features that eventually lead to an output that is the probability that the image contains cat or dog.<\/span><\/p>\n<p style=\"text-align: justify;\"><span style=\"font-family: helvetica, arial, sans-serif; font-size: 12pt;\">What you essentially do is feed one image at a time to this\u00a0<\/span><span style=\"font-family: helvetica, arial, sans-serif; font-size: 12pt;\">network of the training database and ask it to predict the correct label all the time. Eventually all these sort of connections settle onto a configuration that works well and generalizes beyond the images that means being trainer and internally so well that in fact the billions of pictures every single day are being analyzed through this kind of algorithm.<\/span><\/p>\n<p><img decoding=\"async\" loading=\"lazy\" class=\"alignnone wp-image-3622\" src=\"http:\/\/www.artificialbrain.xyz\/wp-content\/uploads\/Google-Machine-Translation-450x253.jpg\" alt=\"\" width=\"784\" height=\"441\" srcset=\"https:\/\/www.newworldai.com\/wp-content\/uploads\/Google-Machine-Translation-450x253.jpg 450w, https:\/\/www.newworldai.com\/wp-content\/uploads\/Google-Machine-Translation-178x100.jpg 178w, https:\/\/www.newworldai.com\/wp-content\/uploads\/Google-Machine-Translation-768x432.jpg 768w, https:\/\/www.newworldai.com\/wp-content\/uploads\/Google-Machine-Translation-1024x576.jpg 1024w, https:\/\/www.newworldai.com\/wp-content\/uploads\/Google-Machine-Translation.jpg 1368w\" sizes=\"(max-width: 784px) 100vw, 784px\" \/><\/p>\n<p style=\"text-align: justify;\"><span style=\"font-family: helvetica, arial, sans-serif; font-size: 12pt;\">The next and one of the sort of breakthroughs in 2014 which me incorporators and other groups are sort of discovered is not only this idea of taking an image and producing a probability but also taking a sort of sequence of symbols and mapping it through learning examples to another sequence of symbols.<\/span><\/p>\n<p style=\"text-align: justify;\"><span style=\"font-family: helvetica, arial, sans-serif; font-size: 12pt;\">This is a very powerful concept. Because, now we can also learn for instance to do machine translation. So, Here the sequence the input sequence would be the Chinese characters and the\u00a0<\/span><span style=\"font-family: helvetica, arial, sans-serif; font-size: 12pt;\">output would be English words. Crucially, you learn these without any knowledge of a language or which words\u00a0<\/span><span style=\"font-family: helvetica, arial, sans-serif; font-size: 12pt;\">correspond to which words. All you need is a large database of pairs of\u00a0<\/span><span style=\"font-family: helvetica, arial, sans-serif; font-size: 12pt;\">sentences in Chinese and in English translations of each other.\u00a0<\/span><\/p>\n<p><img decoding=\"async\" loading=\"lazy\" class=\"alignnone wp-image-3623\" src=\"http:\/\/www.artificialbrain.xyz\/wp-content\/uploads\/Google-Machine-Translations-450x253.jpg\" alt=\"\" width=\"802\" height=\"451\" srcset=\"https:\/\/www.newworldai.com\/wp-content\/uploads\/Google-Machine-Translations-450x253.jpg 450w, https:\/\/www.newworldai.com\/wp-content\/uploads\/Google-Machine-Translations-178x100.jpg 178w, https:\/\/www.newworldai.com\/wp-content\/uploads\/Google-Machine-Translations-768x431.jpg 768w, https:\/\/www.newworldai.com\/wp-content\/uploads\/Google-Machine-Translations-1024x575.jpg 1024w, https:\/\/www.newworldai.com\/wp-content\/uploads\/Google-Machine-Translations.jpg 1372w\" sizes=\"(max-width: 802px) 100vw, 802px\" \/><\/p>\n<p style=\"text-align: justify;\"><span style=\"font-family: helvetica, arial, sans-serif; font-size: 12pt;\">In fact,\u00a0<\/span><span style=\"font-family: helvetica, arial, sans-serif; font-size: 12pt;\">Google Translate has been upgraded recently to use these sort of neural networks approaches that essentially\u00a0<\/span><span style=\"font-family: helvetica, arial, sans-serif; font-size: 12pt;\">reach the gap between the previous best systems and human quality translations by 50% essentially. These again\u00a0<\/span><span style=\"font-family: helvetica, arial, sans-serif; font-size: 12pt;\">is already used and it&#8217;s sort of already out there for everyone to benefit from.<\/span><\/p>\n<p><img decoding=\"async\" loading=\"lazy\" class=\"alignnone wp-image-3624\" src=\"http:\/\/www.artificialbrain.xyz\/wp-content\/uploads\/WaveNet-450x253.jpg\" alt=\"\" width=\"804\" height=\"452\" srcset=\"https:\/\/www.newworldai.com\/wp-content\/uploads\/WaveNet-450x253.jpg 450w, https:\/\/www.newworldai.com\/wp-content\/uploads\/WaveNet-178x100.jpg 178w, https:\/\/www.newworldai.com\/wp-content\/uploads\/WaveNet-768x432.jpg 768w, https:\/\/www.newworldai.com\/wp-content\/uploads\/WaveNet-1024x576.jpg 1024w, https:\/\/www.newworldai.com\/wp-content\/uploads\/WaveNet.jpg 1369w\" sizes=\"(max-width: 804px) 100vw, 804px\" \/><\/p>\n<p style=\"text-align: justify;\"><span style=\"font-family: helvetica, arial, sans-serif; font-size: 12pt;\">Another such example that is sort of interesting is generating speech. We actually do understand how speech is generated. There is a Google Track which essentially is a bunch of tubes that check changing shape. Y<\/span><span style=\"font-family: helvetica, arial, sans-serif; font-size: 12pt;\">ou could\u00a0<\/span><span style=\"font-family: helvetica, arial, sans-serif; font-size: 12pt;\">actually model this and this is one of these parametric approaches that you see\u00a0<\/span><span style=\"font-family: helvetica, arial, sans-serif; font-size: 12pt;\">there on the left. But you can also do is just record this acoustic the audio\u00a0<\/span><span style=\"font-family: helvetica, arial, sans-serif; font-size: 12pt;\">files essentially from many people which will do very complex from file. There&#8217;s like tens of thousands of points per second of speech but again you just let a neural network figure out the sort of patterns in this signal.<\/span><\/p>\n<p style=\"text-align: justify;\"><span style=\"font-family: helvetica, arial, sans-serif; font-size: 12pt;\">What you get end up with is a very good natural-sounding speech system that again beats\u00a0<\/span><span style=\"font-family: helvetica, arial, sans-serif; font-size: 12pt;\">the gap between the previous best and human speech by about 50% factor. T<\/span><span style=\"font-family: helvetica, arial, sans-serif; font-size: 12pt;\">his is some like drastic already sort of well-developed ideas that supervised learning enables us to do and it&#8217;s all about this idea of imitation learning. You imitate humans through these powerful models deep neural Nets.<\/span><\/p>\n<p style=\"text-align: justify;\"><span style=\"font-family: helvetica, arial, sans-serif; font-size: 12pt;\">There is another paradigm I wanted to describe which is what if we cannot get a large database of these examples or even better. We want to actually go beyond human capabilities right if we actually want to beat humans.<\/span><span style=\"font-family: helvetica, arial, sans-serif; font-size: 12pt;\">\u00a0This is the reinforcement learning paradigm and for instance in the game of GO, this was heavily used to for the recent result.<\/span><\/p>\n<p style=\"text-align: justify;\"><span style=\"font-family: helvetica, arial, sans-serif; font-size: 12pt;\">It is extremely simple how this works. It&#8217;s almost like obvious. The first step is we have an agent which is in this case the Go player and this agent lives in an environment. The environment would be the Go world. This agent has a goal. For instance, it wants to win at the game of Go and the way this works is quite simple.<\/span><\/p>\n<p><img decoding=\"async\" loading=\"lazy\" class=\"alignnone wp-image-3625\" src=\"http:\/\/www.artificialbrain.xyz\/wp-content\/uploads\/The-Reinforcement-Learning-Paradigm-450x253.jpg\" alt=\"\" width=\"802\" height=\"451\" srcset=\"https:\/\/www.newworldai.com\/wp-content\/uploads\/The-Reinforcement-Learning-Paradigm-450x253.jpg 450w, https:\/\/www.newworldai.com\/wp-content\/uploads\/The-Reinforcement-Learning-Paradigm-178x100.jpg 178w, https:\/\/www.newworldai.com\/wp-content\/uploads\/The-Reinforcement-Learning-Paradigm-768x432.jpg 768w, https:\/\/www.newworldai.com\/wp-content\/uploads\/The-Reinforcement-Learning-Paradigm-1024x576.jpg 1024w, https:\/\/www.newworldai.com\/wp-content\/uploads\/The-Reinforcement-Learning-Paradigm.jpg 1368w\" sizes=\"(max-width: 802px) 100vw, 802px\" \/><\/p>\n<p style=\"text-align: justify;\"><span style=\"font-family: helvetica, arial, sans-serif; font-size: 12pt;\">The agent observes the environment. <\/span><span style=\"font-family: helvetica, arial, sans-serif; font-size: 12pt;\">When the game finishes and there my environment as \u00a0won the game or not. After it&#8217;s been taught whether or not it updates its ways to ensure that essentially the reward that it gets the probability of winning\u00a0<\/span><span style=\"font-family: helvetica, arial, sans-serif; font-size: 12pt;\">is maximized.<\/span><\/p>\n<p><img decoding=\"async\" loading=\"lazy\" class=\"alignnone wp-image-3626\" src=\"http:\/\/www.artificialbrain.xyz\/wp-content\/uploads\/human-level-control-through-deep-reinforcement-learning-450x254.jpg\" alt=\"\" width=\"803\" height=\"453\" srcset=\"https:\/\/www.newworldai.com\/wp-content\/uploads\/human-level-control-through-deep-reinforcement-learning-450x254.jpg 450w, https:\/\/www.newworldai.com\/wp-content\/uploads\/human-level-control-through-deep-reinforcement-learning-177x100.jpg 177w, https:\/\/www.newworldai.com\/wp-content\/uploads\/human-level-control-through-deep-reinforcement-learning-768x434.jpg 768w, https:\/\/www.newworldai.com\/wp-content\/uploads\/human-level-control-through-deep-reinforcement-learning-1024x578.jpg 1024w, https:\/\/www.newworldai.com\/wp-content\/uploads\/human-level-control-through-deep-reinforcement-learning.jpg 1364w\" sizes=\"(max-width: 803px) 100vw, 803px\" \/><\/p>\n<p style=\"text-align: justify;\"><span style=\"font-family: helvetica, arial, sans-serif; font-size: 12pt;\">This agent is very very similar to the neural network that was classifying\u00a0<\/span><span style=\"font-family: helvetica, arial, sans-serif; font-size: 12pt;\">cat ord dogs. But in this case, it takes as the inputs these observations so it would be a Go board,\u00a0<\/span><span style=\"font-family: helvetica, arial, sans-serif; font-size: 12pt;\">video games cleaner or not. The outputs are simply the probability of each actions or which judge the position\u00a0<\/span><span style=\"font-family: helvetica, arial, sans-serif; font-size: 12pt;\">you want to play and through this process you can achieve superhuman performance in playing.<\/span><\/p>\n<p style=\"text-align: justify;\"><span style=\"font-family: helvetica, arial, sans-serif; font-size: 12pt;\">Very crucially, it is most of these algorithms have been developed for games which is kind of\u00a0<\/span><span style=\"font-family: helvetica, arial, sans-serif; font-size: 12pt;\">funny. Because, as a kid I used to play video games and now I sort of work on making algorithms for video games. I\u00a0<\/span><span style=\"font-family: helvetica, arial, sans-serif; font-size: 12pt;\">don&#8217;t get to play. But despite these profound realizations what you actually get to do when you do through these two video games is being able to scale. Because, as you can imagine this circle of observations and actions initially it&#8217;s not very good.<\/span><\/p>\n<p style=\"text-align: justify;\"><span style=\"font-family: helvetica, arial, sans-serif; font-size: 12pt;\">In other words, if you act randomly, you would not gonna get the works very often. So, you have to play a lot in order to achieve good performance. But if you go out to the real world,\u00a0<\/span><span style=\"font-family: helvetica, arial, sans-serif; font-size: 12pt;\">things will be very slow if you are to build 10,000 robots that you would crash and eventually you learn how to do this that wouldn&#8217;t scale very well.<\/span><\/p>\n<p style=\"text-align: justify;\"><span style=\"font-family: helvetica, arial, sans-serif; font-size: 12pt;\">Video games are very important. Because, you can&#8217;t run them in parallel in a cluster and if you lose it&#8217;s not a big deal. But eventually you are testing sort\u00a0<\/span><span style=\"font-family: helvetica, arial, sans-serif; font-size: 12pt;\">of ideas about intelligence. Through these video guys which were designed by making\u00a0<\/span><span style=\"font-family: helvetica, arial, sans-serif; font-size: 12pt;\">essentially humans, imagine and plan and so on. This is our sort of research platform it has been so far until recently. It&#8217;s gonna be in the future as well.<\/span><\/p>\n<p><img decoding=\"async\" loading=\"lazy\" class=\"alignnone wp-image-3629\" src=\"http:\/\/www.artificialbrain.xyz\/wp-content\/uploads\/environments-450x253.jpg\" alt=\"\" width=\"937\" height=\"527\" srcset=\"https:\/\/www.newworldai.com\/wp-content\/uploads\/environments-450x253.jpg 450w, https:\/\/www.newworldai.com\/wp-content\/uploads\/environments-178x100.jpg 178w, https:\/\/www.newworldai.com\/wp-content\/uploads\/environments-768x432.jpg 768w, https:\/\/www.newworldai.com\/wp-content\/uploads\/environments-1024x576.jpg 1024w, https:\/\/www.newworldai.com\/wp-content\/uploads\/environments.jpg 1368w\" sizes=\"(max-width: 937px) 100vw, 937px\" \/><\/p>\n<p style=\"text-align: justify;\"><span style=\"font-family: helvetica, arial, sans-serif; font-size: 12pt;\">But what&#8217;s the problem well the environment if the environment is very simple we will learn by simple agents so the environment and the agents\u00a0<\/span><span style=\"font-family: helvetica, arial, sans-serif; font-size: 12pt;\">go side by side and as you can imagine the is extremely complex. There are like\u00a0<\/span><span style=\"font-family: helvetica, arial, sans-serif; font-size: 12pt;\">you know predators out there that might want to eat you you need to energy so you might want to eat plants or other other human being not human beings beings and and it&#8217;s very diverse and very complex<\/span><span style=\"font-family: helvetica, arial, sans-serif; font-size: 12pt;\">.<\/span><\/p>\n<p><img decoding=\"async\" loading=\"lazy\" class=\"alignnone wp-image-3630\" src=\"http:\/\/www.artificialbrain.xyz\/wp-content\/uploads\/simulated-environments-450x252.jpg\" alt=\"\" width=\"804\" height=\"450\" srcset=\"https:\/\/www.newworldai.com\/wp-content\/uploads\/simulated-environments-450x252.jpg 450w, https:\/\/www.newworldai.com\/wp-content\/uploads\/simulated-environments-178x100.jpg 178w, https:\/\/www.newworldai.com\/wp-content\/uploads\/simulated-environments-768x430.jpg 768w, https:\/\/www.newworldai.com\/wp-content\/uploads\/simulated-environments-1024x574.jpg 1024w, https:\/\/www.newworldai.com\/wp-content\/uploads\/simulated-environments-715x400.jpg 715w, https:\/\/www.newworldai.com\/wp-content\/uploads\/simulated-environments.jpg 1374w\" sizes=\"(max-width: 804px) 100vw, 804px\" \/><\/p>\n<p style=\"text-align: justify;\"><span style=\"font-family: helvetica, arial, sans-serif; font-size: 12pt;\">It seems like\u00a0<\/span><span style=\"font-family: helvetica, arial, sans-serif; font-size: 12pt;\">creating environments and creating agents goes both hand by hand and he didn&#8217;t need that so one of the the ideas\u00a0<\/span><span style=\"font-family: helvetica, arial, sans-serif; font-size: 12pt;\">here is if we are simulating the the natural world and we want for instance to get this guy to like survive the\u00a0<\/span><span style=\"font-family: helvetica, arial, sans-serif; font-size: 12pt;\">winter we we need to simulate snow and food and so on and that is a hard problem by itself as well.<\/span><\/p>\n<p style=\"text-align: justify;\"><span style=\"font-family: helvetica, arial, sans-serif; font-size: 12pt;\">However you\u00a0<\/span><span style=\"font-family: helvetica, arial, sans-serif; font-size: 12pt;\">know what you can actually try to do is well we researchers not only build agents but we are actually also seeing\u00a0<\/span><span style=\"font-family: helvetica, arial, sans-serif; font-size: 12pt;\">about the environments that will make agents advance and amongst many new environments that people create last\u00a0<\/span><span style=\"font-family: helvetica, arial, sans-serif; font-size: 12pt;\">year we created this environment which is called the mine lab which looks like three dimensional you know researchers can set all sorts of tasks to test to test ideas about planning memory\u00a0<\/span><span style=\"font-family: helvetica, arial, sans-serif; font-size: 12pt;\">imagination and so on.<\/span><\/p>\n<p><img decoding=\"async\" loading=\"lazy\" class=\"alignnone wp-image-3631\" src=\"http:\/\/www.artificialbrain.xyz\/wp-content\/uploads\/DeepMind-Lab-450x253.jpg\" alt=\"\" width=\"809\" height=\"455\" srcset=\"https:\/\/www.newworldai.com\/wp-content\/uploads\/DeepMind-Lab-450x253.jpg 450w, https:\/\/www.newworldai.com\/wp-content\/uploads\/DeepMind-Lab-178x100.jpg 178w, https:\/\/www.newworldai.com\/wp-content\/uploads\/DeepMind-Lab-768x432.jpg 768w, https:\/\/www.newworldai.com\/wp-content\/uploads\/DeepMind-Lab-1024x576.jpg 1024w, https:\/\/www.newworldai.com\/wp-content\/uploads\/DeepMind-Lab.jpg 1368w\" sizes=\"(max-width: 809px) 100vw, 809px\" \/><\/p>\n<p style=\"text-align: justify;\"><span style=\"font-family: helvetica, arial, sans-serif; font-size: 12pt;\">These environments are becoming more complex and as a result our agents will also of course learn better there.<\/span><\/p>\n<p><img decoding=\"async\" loading=\"lazy\" class=\"alignnone wp-image-3632\" src=\"http:\/\/www.artificialbrain.xyz\/wp-content\/uploads\/AI-needs-nature-450x254.jpg\" alt=\"\" width=\"804\" height=\"454\" srcset=\"https:\/\/www.newworldai.com\/wp-content\/uploads\/AI-needs-nature-450x254.jpg 450w, https:\/\/www.newworldai.com\/wp-content\/uploads\/AI-needs-nature-177x100.jpg 177w, https:\/\/www.newworldai.com\/wp-content\/uploads\/AI-needs-nature-768x434.jpg 768w, https:\/\/www.newworldai.com\/wp-content\/uploads\/AI-needs-nature-1024x578.jpg 1024w, https:\/\/www.newworldai.com\/wp-content\/uploads\/AI-needs-nature.jpg 1364w\" sizes=\"(max-width: 804px) 100vw, 804px\" \/><\/p>\n<p style=\"text-align: justify;\"><span style=\"font-family: helvetica, arial, sans-serif; font-size: 12pt;\">It is kind of obvious that we need artificial intelligence. It&#8217;s very helpful to be able to have\u00a0<\/span><span style=\"font-family: helvetica, arial, sans-serif; font-size: 12pt;\">translation systems, you might have a personal assistant on your phone like helping you organize streets and so on so forth. <\/span><\/p>\n<p style=\"text-align: justify;\"><span style=\"font-family: helvetica, arial, sans-serif; font-size: 12pt;\">But it also seems that AI needs nature we need to encode these complex rules for AI to be useful to us. Otherwise, we are just bound by the complexity of the environment in which agents live in.<\/span><\/p>\n<p style=\"text-align: justify;\"><img decoding=\"async\" loading=\"lazy\" class=\"alignnone wp-image-3620\" src=\"http:\/\/www.artificialbrain.xyz\/wp-content\/uploads\/Oriol-Vinyals-450x270.png\" alt=\"\" width=\"803\" height=\"482\" srcset=\"https:\/\/www.newworldai.com\/wp-content\/uploads\/Oriol-Vinyals-450x270.png 450w, https:\/\/www.newworldai.com\/wp-content\/uploads\/Oriol-Vinyals-166x100.png 166w, https:\/\/www.newworldai.com\/wp-content\/uploads\/Oriol-Vinyals-768x461.png 768w, https:\/\/www.newworldai.com\/wp-content\/uploads\/Oriol-Vinyals-1024x615.png 1024w, https:\/\/www.newworldai.com\/wp-content\/uploads\/Oriol-Vinyals-184x109.png 184w, https:\/\/www.newworldai.com\/wp-content\/uploads\/Oriol-Vinyals.png 1368w\" sizes=\"(max-width: 803px) 100vw, 803px\" \/><\/p>\n<p style=\"text-align: justify;\"><span style=\"font-family: helvetica, arial, sans-serif; font-size: 12pt;\">Oriol Vinyals is a Research Scientist at Google DeepMind, working on Deep Learning. Oriol holds a Ph.D. in EECS from University of California, Berkeley, a Masters degree from University of California, San Diego, and a double degree in Mathematics and Telecommunication Engineering from UPC, Barcelona. At Google Brain and Google DeepMind he continues working on his areas of interest, which include artificial intelligence, with particular emphasis on machine learning, language, and vision.<\/span><\/p>\n<p style=\"text-align: justify;\"><span style=\"font-family: helvetica, arial, sans-serif; font-size: 12pt;\">Oriol Vinyals is a Research Scientist at Google DeepMind, working on Deep Learning. Oriol holds a Ph.D. in EECS from University of California, Berkeley, a Masters degree from University of California, San Diego, and a double degree in Mathematics and Telecommunication Engineering from UPC, Barcelona. He is a recipient of the 2011 Microsoft Research PhD Fellowship. He was an early adopter of the new deep learning wave at Berkeley, and in his thesis he focused on non-convex optimization and recurrent neural networks. At Google Brain and Google DeepMind he continues working on his areas of interest, which include artificial intelligence, with particular emphasis on machine learning, language, and vision.<\/span><\/p>\n","protected":false},"excerpt":{"rendered":"<p>What is artificial general intelligence and what are researchers doing to achieve this goal today? Oriol Vinyals, from Google DeepMind, walks you through the<\/p>\n","protected":false},"author":2,"featured_media":3627,"comment_status":"open","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":[],"categories":[5],"tags":[325,459,18,180,976],"yoast_head":"<!-- This site is optimized with the Yoast SEO plugin v20.2.1 - https:\/\/yoast.com\/wordpress\/plugins\/seo\/ -->\n<title>Towards Artificial General Intelligence: Oriol Vinyals - New World : Artificial Intelligence<\/title>\n<meta name=\"robots\" content=\"index, follow, max-snippet:-1, max-image-preview:large, max-video-preview:-1\" \/>\n<link rel=\"canonical\" href=\"https:\/\/www.newworldai.com\/towards-artificial-general-intelligence-oriol-vinyals\/\" \/>\n<meta property=\"og:locale\" content=\"en_US\" \/>\n<meta property=\"og:type\" content=\"article\" \/>\n<meta property=\"og:title\" content=\"Towards Artificial General Intelligence: Oriol Vinyals - New World : Artificial Intelligence\" \/>\n<meta property=\"og:description\" content=\"What is artificial general intelligence and what are researchers doing to achieve this goal today? Oriol Vinyals, from Google DeepMind, walks you through the\" \/>\n<meta property=\"og:url\" content=\"https:\/\/www.newworldai.com\/towards-artificial-general-intelligence-oriol-vinyals\/\" \/>\n<meta property=\"og:site_name\" content=\"New World : Artificial Intelligence\" \/>\n<meta property=\"article:publisher\" content=\"https:\/\/www.facebook.com\/groups\/952169468169594\/\" \/>\n<meta property=\"article:published_time\" content=\"2023-01-09T21:59:12+00:00\" \/>\n<meta property=\"article:modified_time\" content=\"2023-01-12T22:17:40+00:00\" \/>\n<meta property=\"og:image\" content=\"https:\/\/www.newworldai.com\/wp-content\/uploads\/Towards-Artificial-General-Intelligence-Oriol-Vinyals.jpg\" \/>\n\t<meta property=\"og:image:width\" content=\"1368\" \/>\n\t<meta property=\"og:image:height\" content=\"770\" \/>\n\t<meta property=\"og:image:type\" content=\"image\/jpeg\" \/>\n<meta name=\"author\" content=\"mussem\" \/>\n<meta name=\"twitter:card\" content=\"summary_large_image\" \/>\n<meta name=\"twitter:creator\" content=\"@NewWorldAI1\" \/>\n<meta name=\"twitter:site\" content=\"@NewWorldAI1\" \/>\n<meta name=\"twitter:label1\" content=\"Written by\" \/>\n\t<meta name=\"twitter:data1\" content=\"mussem\" \/>\n\t<meta name=\"twitter:label2\" content=\"Est. reading time\" \/>\n\t<meta name=\"twitter:data2\" content=\"11 minutes\" \/>\n<script type=\"application\/ld+json\" class=\"yoast-schema-graph\">{\"@context\":\"https:\/\/schema.org\",\"@graph\":[{\"@type\":\"WebPage\",\"@id\":\"https:\/\/www.newworldai.com\/towards-artificial-general-intelligence-oriol-vinyals\/\",\"url\":\"https:\/\/www.newworldai.com\/towards-artificial-general-intelligence-oriol-vinyals\/\",\"name\":\"Towards Artificial General Intelligence: Oriol Vinyals - New World : Artificial Intelligence\",\"isPartOf\":{\"@id\":\"https:\/\/www.newworldai.com\/#website\"},\"datePublished\":\"2023-01-09T21:59:12+00:00\",\"dateModified\":\"2023-01-12T22:17:40+00:00\",\"author\":{\"@id\":\"https:\/\/www.newworldai.com\/#\/schema\/person\/9d8dd21ba4efacdbf5a5a7700ac3ec1f\"},\"breadcrumb\":{\"@id\":\"https:\/\/www.newworldai.com\/towards-artificial-general-intelligence-oriol-vinyals\/#breadcrumb\"},\"inLanguage\":\"en-US\",\"potentialAction\":[{\"@type\":\"ReadAction\",\"target\":[\"https:\/\/www.newworldai.com\/towards-artificial-general-intelligence-oriol-vinyals\/\"]}]},{\"@type\":\"BreadcrumbList\",\"@id\":\"https:\/\/www.newworldai.com\/towards-artificial-general-intelligence-oriol-vinyals\/#breadcrumb\",\"itemListElement\":[{\"@type\":\"ListItem\",\"position\":1,\"name\":\"Home\",\"item\":\"https:\/\/www.newworldai.com\/\"},{\"@type\":\"ListItem\",\"position\":2,\"name\":\"Towards Artificial General Intelligence: Oriol Vinyals\"}]},{\"@type\":\"WebSite\",\"@id\":\"https:\/\/www.newworldai.com\/#website\",\"url\":\"https:\/\/www.newworldai.com\/\",\"name\":\"New World : Artificial Intelligence\",\"description\":\"Artificial Intelligence, Deep Learning, Machine Learning, AI Lectures, AI Conferences, AI TED Talks, AI Movies, AI Books\",\"potentialAction\":[{\"@type\":\"SearchAction\",\"target\":{\"@type\":\"EntryPoint\",\"urlTemplate\":\"https:\/\/www.newworldai.com\/?s={search_term_string}\"},\"query-input\":\"required name=search_term_string\"}],\"inLanguage\":\"en-US\"},{\"@type\":\"Person\",\"@id\":\"https:\/\/www.newworldai.com\/#\/schema\/person\/9d8dd21ba4efacdbf5a5a7700ac3ec1f\",\"name\":\"mussem\",\"image\":{\"@type\":\"ImageObject\",\"inLanguage\":\"en-US\",\"@id\":\"https:\/\/www.newworldai.com\/#\/schema\/person\/image\/\",\"url\":\"https:\/\/secure.gravatar.com\/avatar\/b2359d117d746a042c3d719211039fe2?s=96&d=mm&r=g\",\"contentUrl\":\"https:\/\/secure.gravatar.com\/avatar\/b2359d117d746a042c3d719211039fe2?s=96&d=mm&r=g\",\"caption\":\"mussem\"}}]}<\/script>\n<!-- \/ Yoast SEO plugin. -->","yoast_head_json":{"title":"Towards Artificial General Intelligence: Oriol Vinyals - New World : Artificial Intelligence","robots":{"index":"index","follow":"follow","max-snippet":"max-snippet:-1","max-image-preview":"max-image-preview:large","max-video-preview":"max-video-preview:-1"},"canonical":"https:\/\/www.newworldai.com\/towards-artificial-general-intelligence-oriol-vinyals\/","og_locale":"en_US","og_type":"article","og_title":"Towards Artificial General Intelligence: Oriol Vinyals - New World : Artificial Intelligence","og_description":"What is artificial general intelligence and what are researchers doing to achieve this goal today? Oriol Vinyals, from Google DeepMind, walks you through the","og_url":"https:\/\/www.newworldai.com\/towards-artificial-general-intelligence-oriol-vinyals\/","og_site_name":"New World : Artificial Intelligence","article_publisher":"https:\/\/www.facebook.com\/groups\/952169468169594\/","article_published_time":"2023-01-09T21:59:12+00:00","article_modified_time":"2023-01-12T22:17:40+00:00","og_image":[{"width":1368,"height":770,"url":"https:\/\/www.newworldai.com\/wp-content\/uploads\/Towards-Artificial-General-Intelligence-Oriol-Vinyals.jpg","type":"image\/jpeg"}],"author":"mussem","twitter_card":"summary_large_image","twitter_creator":"@NewWorldAI1","twitter_site":"@NewWorldAI1","twitter_misc":{"Written by":"mussem","Est. reading time":"11 minutes"},"schema":{"@context":"https:\/\/schema.org","@graph":[{"@type":"WebPage","@id":"https:\/\/www.newworldai.com\/towards-artificial-general-intelligence-oriol-vinyals\/","url":"https:\/\/www.newworldai.com\/towards-artificial-general-intelligence-oriol-vinyals\/","name":"Towards Artificial General Intelligence: Oriol Vinyals - New World : Artificial Intelligence","isPartOf":{"@id":"https:\/\/www.newworldai.com\/#website"},"datePublished":"2023-01-09T21:59:12+00:00","dateModified":"2023-01-12T22:17:40+00:00","author":{"@id":"https:\/\/www.newworldai.com\/#\/schema\/person\/9d8dd21ba4efacdbf5a5a7700ac3ec1f"},"breadcrumb":{"@id":"https:\/\/www.newworldai.com\/towards-artificial-general-intelligence-oriol-vinyals\/#breadcrumb"},"inLanguage":"en-US","potentialAction":[{"@type":"ReadAction","target":["https:\/\/www.newworldai.com\/towards-artificial-general-intelligence-oriol-vinyals\/"]}]},{"@type":"BreadcrumbList","@id":"https:\/\/www.newworldai.com\/towards-artificial-general-intelligence-oriol-vinyals\/#breadcrumb","itemListElement":[{"@type":"ListItem","position":1,"name":"Home","item":"https:\/\/www.newworldai.com\/"},{"@type":"ListItem","position":2,"name":"Towards Artificial General Intelligence: Oriol Vinyals"}]},{"@type":"WebSite","@id":"https:\/\/www.newworldai.com\/#website","url":"https:\/\/www.newworldai.com\/","name":"New World : Artificial Intelligence","description":"Artificial Intelligence, Deep Learning, Machine Learning, AI Lectures, AI Conferences, AI TED Talks, AI Movies, AI Books","potentialAction":[{"@type":"SearchAction","target":{"@type":"EntryPoint","urlTemplate":"https:\/\/www.newworldai.com\/?s={search_term_string}"},"query-input":"required name=search_term_string"}],"inLanguage":"en-US"},{"@type":"Person","@id":"https:\/\/www.newworldai.com\/#\/schema\/person\/9d8dd21ba4efacdbf5a5a7700ac3ec1f","name":"mussem","image":{"@type":"ImageObject","inLanguage":"en-US","@id":"https:\/\/www.newworldai.com\/#\/schema\/person\/image\/","url":"https:\/\/secure.gravatar.com\/avatar\/b2359d117d746a042c3d719211039fe2?s=96&d=mm&r=g","contentUrl":"https:\/\/secure.gravatar.com\/avatar\/b2359d117d746a042c3d719211039fe2?s=96&d=mm&r=g","caption":"mussem"}}]}},"_links":{"self":[{"href":"https:\/\/www.newworldai.com\/wp-json\/wp\/v2\/posts\/3571"}],"collection":[{"href":"https:\/\/www.newworldai.com\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.newworldai.com\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.newworldai.com\/wp-json\/wp\/v2\/users\/2"}],"replies":[{"embeddable":true,"href":"https:\/\/www.newworldai.com\/wp-json\/wp\/v2\/comments?post=3571"}],"version-history":[{"count":2,"href":"https:\/\/www.newworldai.com\/wp-json\/wp\/v2\/posts\/3571\/revisions"}],"predecessor-version":[{"id":5504,"href":"https:\/\/www.newworldai.com\/wp-json\/wp\/v2\/posts\/3571\/revisions\/5504"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/www.newworldai.com\/wp-json\/wp\/v2\/media\/3627"}],"wp:attachment":[{"href":"https:\/\/www.newworldai.com\/wp-json\/wp\/v2\/media?parent=3571"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.newworldai.com\/wp-json\/wp\/v2\/categories?post=3571"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.newworldai.com\/wp-json\/wp\/v2\/tags?post=3571"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}