Jupyter Notebook,Dind,Ubuntu-16.04,Anaconda3,Tensorflow-1.8,OpenCV-3,OpenAI-gym-universe
10K+
I think it’s important to point out that OpenAI-gym-universe with opencv3 are difficult to set-up properly. This may be due to the lag of proper intro of OpenAI-gym installation guide on the older version and hopefully this docker repo would help to kick-start.
There is actually a way to install OpenAI-gym-universe on JupyterNotebook using conda (easier method). Somehow most of the articles I found online describe the process of building from source code, which is painfully long and error-prone. Hence this repo may ease all the pain and hazard during set-up on :
(Cannot run at PWD due run --privileged violate security policy OK at local computer)
OpenAI is a non-profit artificial intelligence (AI) research company that aims to promote and develop friendly AI in such a way as to benefit humanity as a whole. The organization aims to "freely collaborate" with other institutions and researchers by making its patents and research open to the public.The founders (notably Elon Musk and Sam Altman) are motivated in part by concerns about existential risk from artificial general intelligence. Game programmers used to use heuristic if-then-else type decisions to make educated guesses. We saw this in the earliest arcade videos games such as Pong and PacMan. This trend was the norm for a very long time. But game developers can only predict so many scenarios and edge cases so your bot doesn’t run in circles! Game developers then tried to mimic how humans would play a game, and modeled human intelligence in a game bot. The team at DeepMind did this by generalizing and modeling intelligence to solve any Atari game thrown at it. The game bot used deep learning neural networks that would have no game-specific knowledge. They beat the game based on the pixels they saw on screen and their knowledge of the game controls. However, parts of DeepMind are still not open-sourced as Google uses it to beat competition.
OpenAI, a non-profit artificial intelligence research company, is well-known in the industry. On December 12, 2015, Elon Musk announced on Twitter that it officially launched OpenAI, a non-profit artificial intelligence project. The goal of OpenAI is to "drive the development of digital intelligence without being limited by financial returns, thereby benefiting the entire human race." Open is the meaning of "open" in English, and AI is an abbreviation of Artificial Intelligence. As the name suggests, OpenAI's vision is to open AI technology to all human beings. OpenAI strongly recommends that researchers disclose their research, including papers, blog posts, code, and patents, to share with the world. Elon Musk is one of the co-founders of OpenAI and the CEO and product architect of Tesla (the company that builds electric cars, now in many other areas), and space exploration technology company (SpaceX, The company that launched a lot of rockets, the CEO and CTO, is also one of the founders of PayPal. It is a big man in the tech world. Anyway, it’s the right time to start a business. Humans can’t stop Musk.
Recently, OpenAI announced that Elon Musk announced its withdrawal from the board of directors because Tesla is more focused on AI technology and needs to avoid possible future conflicts of interest with OpenAI. Although it doesn't explain the specific "points of conflict", it can be thought of that, for example, talent competition, OpenAI began to dig people from technology giants such as Google and Facebook. Now Tesla pays attention to AI technology. The flow of talent is inevitably subject to criticism. OpenAI researcher Andrej Kapathy is great AI specialist. However, according to the announcement, most of the exit from the board of directors is Musk to avoid suspicion. In fact, he still has half a foot left in OpenAI. Musk will continue to serve as an OpenAI consultant and continue to provide funding and advice.
It is said that one of the main reasons why Elon Musk founded OpenAI is because he is the person who holds the "AI threat theory" (similar to him, Dr. Stephen Hawking, etc.. He has the opposite opinion ("AI useful theory") There is Facebook CEO Mark Zuckerberg). Elon Musk felt it necessary to let people know about AI (artificial intelligence) to prevent the possible "AI annihilation" in the future. OpenAI has made a lot of contributions in the field of artificial intelligence, such as the development of Dota 2 in the summer of 2017 (I used to play a lot of Dota 1 in college, when I went to the room with the roommates to go black) It's very interesting. Dota 2 I haven't played since, because I didn't play games and then I went to learn programming. Now I think programming is more fun than playing games, especially now that artificial intelligence programming can also be used to let AI help you play games.) The AI program that beats the strongest human player:
Report on this sensational big news (with video) at https://blog.openai.com/dota-2/ At that time, the world's strongest shadow fan (Shadow Fiend, referred to as SF, has always been a popular hero in Dota, is a hero who tests the operation), one of the users of the Dendi operation, the magical magic program developed by OpenAI in the middle road 1v1 The operation of the Shadow Magic got a blood (First Blood), two blood, three blood... Amount, directly admit defeat (Dendi inner OS: I did not expect the baby to be abused...). Before the game, Dendi still had a relatively relaxed attitude. As a result, the artificial intelligence program was very clever (but on the one hand, because the artificial intelligence program knows all the elements of the game, including the distance, it can release the "three consecutive pressures" of the shadow magic with great precision.) He was greatly surprised. Moreover, the artificial intelligence control of the shadow magic will be the card position, long-distance pursuit, replenishment, killing the other party to transport small animals, pretending to raise hands and cancel the attack, and even made a stunning move such as the top tower.
OpenAI focuses on two key points in machine learning: Unsupervised Learning and Reinforcement Learning. Some of its important projects are basically open source on Github: https://github.com/openai such as: Gym: This article will be introduced later. Https://github.com/openai/gym MuJoCo-py : Use with Python MuJoCo is the physics engine. Https://github.com/openai/mujoco-py RoboSchool: OpenAI's open source robot simulation software. Based on Gym. Https://blog.openai.com/roboschool Baselines: A classic (deep) reinforcement learning algorithm based on Gym and TensorFlow. Https://github.com/openai/baselines Universe: Https://github.com/openai/universe
OpenAI Gym is a toolkit for developing and comparing RL (Reinforcement Learning) algorithms. It includes a growing and streamlined environment and a platform for comparing and evaluating algorithms. Gym is compatible with other numerical computing libraries such as TensorFlow or Theano. The main support is the Python language. The official website of Gym is https://github.com/openai/gym The various environments of Gym are here https://gym.openai.com/envs A good thing about Gym is that you can upload algorithms for your own training environment, you can also see algorithms uploaded by others, and you can download other people's algorithm code. For example, https://gym.openai.com/envs/CartPole-v0 is the environment for CartPole. The purpose is to set up the pole, where you can find algorithms uploaded by others: Algorithm uploaded by others Based on Gym, Universe is a software platform for evaluating and training intelligent agents in games, web pages and other applications around the world. The agent uses sensory input and control similar to humans, but it sees pixels and controls the mouse and keyboard.
This allows any AI to perform tasks that can be trained to do AI and compete with human players. The Universe contains more than 1,000 different training environments, including games like Flash games, web tasks, and Grand Theft Auto. Universe is the meaning of the English word "universe". It can be seen that the vision of this project is very large, aiming to develop an algorithm close to General Artificial Intelligence. The official website of Universe is https://github.com/openai/universe (originally https://universe.openai.com is the official website, but not long ago this website has been turned to the page on Github, it seems that the Universe project Not as active as the Gym project). But Universe also has a blog: https://blog.openai.com/universe The article inside is worth a look and introduces the basic principles of the universe. The main point is that each environment of the universe is basically running in a Docker container, then connected remotely through VNC, etc.
.................................................................................................................................................................... Please drop me email/whatever reason? [email protected]@gmail.com for discussion to advance AI!
Content type
Image
Digest
Size
2.8 GB
Last updated
about 8 years ago
docker pull mltandocker/dind-universe