It aims to promote and develop friendly AI in a way to benefit humanity as a whole. It aims to "freely collaborate" by making its patents and research open to the public.

To put that into perspective, experts at OpenAI recently developed Dactyl, a robotic hand that could handle objects. This is a task that any human child learns to perform subconsciously at an early age. But it took Dactyl 6,144 CPUs and 8 GPUs and about one hundred years' worth of experience to develop the same skills. While it is a fascinating achievement, it also highlights the stark differences between narrow AI and the way the human brain works.

For engineers at the Elon Musk-founded nonprofit OpenAI, this presented both a challenge and an opportunity. How could their researchers use artificial intelligence to teach a robot to manipulate objects as artfully as a human? Usually, when teaching an AI to control a physical robot, scientists tend to come up against the same problems. Training is often done using reinforcement learning; a method where the AI learns through a process of trial and error. But this requires a lot of time, usually amounting to years of experience.

"OpenAI Five plays 180 years worth of games against itself every day." Artificial Intelligence vs Humans is a hotly debated topic these days and the gaming arena holds one of the prime witnesses of that. DOTA 2 players have taken on an AI algorithm before in a head-to-head match and have lost. The story seems to continue to date, a recent match between a team of 5 humans vs a mix of AI programs, being a proof of that. The AI programs, developed by OpenAI - an AI research lab founded by Elon Musk and Y Combinator president Sam Altman, won 2 out of 3 matches against the 5 semi-professional humans working together as a team.

Summarizing, there are four desiderata that one would like to see in A.I. systems, if they have to compare well with human (or just animal) brains: meta-learning, learning by demonstration ("few-shot learning"), transfer learning and multi-task learning. Meta-learning is particularly relevant in the case of reinforcement learning. It is obvious that reinforcement learning is highly unnatural. DeepMind's AlphaGo and OpenAi Five need to learn from scratch via a huge number of trials. Animals, instead, use built-in or acquired "meta-skills" to learn new tasks in just a few trials. Modern computational theory of meta-learning (learning how to learn) dates back at least to the 1990s, when Schmidhuber published the manifesto "Simple Principles of Metalearning" (1996), followed by his student Sepp Hochreiter ("Learning to Learn Using Gradient Descent", 2001), and by Nicolas Schweighofer and Kenji Doya at Japan's ATR ("Meta-learning in Reinforcement Learning", 2001). Examples of "deep" meta-learning systems of the new generation are: RL Square by Pieter Abbeel's student Yan Duan at UC Berkeley, based on Schulman's TRPO ("RL Square: Fast Reinforcement Learning via Slow Reinforcement Learning", 2016); the "model-agnostic meta-learning" (MAML) of Sergey Levine's student Chelsea Finn at UC Berkeley ("Model-Agnostic Meta-Learning for Fast Adaptation of Deep Networks", 2017); Marcel Binz's thesis at KTH Royal Institute of Technology ("Learning Goal-Directed Behaviour", 2017); Jane Wang's "deep meta-reinforcement learning" at DeepMind ("Learning to Reinforcement Learn", 2017); and OpenAI's Reptile, developed by Alex Nichol and John Schulman, a generalization of Finn's MAML ("On First-Order Meta-Learning Algorithms", 2018). DeepMind's neuroscientist Matthew Botvinick believes that the latter could be a model for how our brain learns: the dopamine system trains another part of the brain, the prefrontal cortex, to operate as its own free-standing learning system ("Prefrontal Cortex as a Meta-reinforcement Learning System", 2018).

This document aims to provide a review on learning with deep generative models (DGMs), which is an highly-active area in machine learning and more generally, artificial intelligence. This review is not meant to be a tutorial, but when necessary, we provide self-contained derivations for completeness. This review has two features. First, though there are different perspectives to classify DGMs, we choose to organize this review from the perspective of graphical modeling, because the learning methods for directed DGMs and undirected DGMs are fundamentally different. Second, we differentiate model definitions from model learning algorithms, since different learning algorithms can be applied to solve the learning problem on the same model, and an algorithm can be applied to learn different models. We thus separate model definition and model learning, with more emphasis on reviewing, differentiating and connecting different learning algorithms. We also discuss promising future research directions. This review is by no means comprehensive as the field is evolving rapidly. The authors apologize in advance for any missed papers and inaccuracies in descriptions. Corrections and comments are highly welcome.

Last Sunday, a particularly unusual DotA 2 tournament took place. DotA, a complicated, real-time strategy game, is among the most popular e-sports in the world. The five players of one team--Blitz, Cap, Fogged, Merlini, and MoonMeander--were ranked in the 99.95th percentile, inarguably among the best DotA 2 players in the world. However, their opponent still defeated them in two out three games, winning the tournament. An evenly matched game is supposed to take 45 minutes, but these two were over in 14 and 21 minutes, respectively. Their opponent was a team of five neural networks developed by Elon Musk's OpenAI, collectively referred to as OpenAI Five.

The human team – made up of popular Twitch streamers and former professionals ranked in the 99.95th percentile – hunkered down to play against the bots known as OpenAI Five in San Francisco on Sunday. OpenAI Five smashed its opponents, winning comfortably in two out of three games. It did lose one game, however, after spectators watching the match live and on Twitch were allowed to pick the pool of heroes – the playable characters in the game. Each hero comes with its own strengths and weaknesses and picking a balanced combination is paramount to winning. If you have too many characters for the same role, the other team will steamroll you.

Later this month, the best Dota 2 teams in the world will meet in Vancouver for the biggest tournament of the year, The International. The annual contest consistently boasts the highest prize pool in eSports (it's up to $23.5 million already this year), not to mention the glory that comes with winning the prestigious event. It may not be long, however, before a team of non-human players becomes worthy of such success. This weekend, the all-bot roster of OpenAI Five took on a team of Dota 2 casters and ex-pro players that individually rank amongst some of the best in the world. OpenAI Five won the best-of-three exhibition match convincingly, and the only reason the human team took a game was thanks to a little help from the audience.

OpenAI's machine learning bots have beaten another team of semi-professionals in Dota 2, in their second public match in the traditional five-versus-five settings. You can watch the action on Twitch – complete with commenters typing in SKYNET! The human team – made up of popular Twitch streamers and former professionals ranked in the 99.95th percentile – hunkered down to play against the bots known as OpenAI Five in San Francisco on Sunday. OpenAI Five smashed its opponents, winning comfortably in two out of three games. It did lose one game, however, after spectators watching the match live and on Twitch were allowed to pick the pool of heroes – the playable characters in the game.