We find that, just as a large transformer model trained on language can generate coherent text, the same exact model trained on pixel sequences can generate coherent image completions and samples.
We’re introducing Jukebox, a neural net that generates music, including rudimentary singing, as raw audio in a variety of genres and artist styles. We’re releasing the model weights and code, along with a tool to explore the generated samples.
We're releasing Safety Gym, a suite of environments and tools for measuring progress towards reinforcement learning agents which respect safety constraints while training.
We've trained a pair of neural networks to solve the Rubik’s Cube with a human-like robot hand.
We've observed agents discovering progressively more complex tool use while playing a simple game of hide-and-seek.
We’ve created MuseNet, a deep neural network that can generate 4-minute musical compositions with 10 different instruments, and can combine styles from country to Mozart to the Beatles.
OpenAI Five is the first AI to beat the world champions in an esports game, having won two back-to-back games versus the world champion Dota 2 team, OG, at Finals this weekend.
We've created OpenAI LP, a new "capped-profit" company that allows us to rapidly increase our investments in compute and talent while including checks and balances to actualize our mission.
We’ve created activation atlases (in collaboration with researchers from Google Brain), a new technique for visualizing interactions between neurons.
We’ve trained a large-scale unsupervised language model which generates coherent paragraphs of text, achieves state-of-the-art performance on many language modeling benchmarks, and performs rudimentary reading comprehension, machine translation, question answering, and summarization.
We've discovered that the gradient noise scale, a simple statistical metric, predicts the parallelizability of neural network training on a wide range of tasks.
We’re releasing Spinning Up in Deep RL, an educational resource designed to let anyone learn to become a skilled practitioner in deep reinforcement learning. Spinning Up consists of crystal-clear examples of RL code, educational exercises, documentation, and tutorials.
We've trained a human-like robot hand to manipulate physical objects with unprecedented dexterity.
Our team of five neural networks, OpenAI Five, has started to defeat amateur human teams at Dota 2.
We've obtained state-of-the-art results on a suite of diverse language tasks with a scalable, task-agnostic system, which we're also releasing. Our approach is a combination of two existing ideas: transformers and unsupervised pre-training.
We're releasing a charter that describes the principles we use to execute on OpenAI's mission. This document reflects the strategy we've refined over the past two years, including feedback from many people internal and external to OpenAI.
We're launching a transfer learning contest that measures a reinforcement learning algorithm's ability to generalize from previous experience.
We're releasing eight simulated robotics environments and a Baselines implementation of Hindsight Experience Replay, all developed for our research over the past year. We've used these environments to train models which work on physical robots.
We’re releasing highly-optimized GPU kernels for an underexplored class of neural network architectures: networks with block-sparse weights. Depending on the chosen sparsity, these kernels can run orders of magnitude faster than cuBLAS or cuSPARSE.
We've found that self-play allows simulated AIs to discover physical skills like tackling, ducking, faking, kicking, catching, and diving for the ball, without explicitly designing an environment with these skills in mind.
We're releasing two new OpenAI Baselines implementations: ACKTR and A2C. A2C is a synchronous, deterministic variant of Asynchronous Advantage Actor Critic (A3C) which we've found gives equal performance.
We've found that adding adaptive noise to the parameters of reinforcement learning algorithms frequently boosts performance. This exploration method is simple to implement and very rarely decreases performance, so it's worth trying on any problem.
We’re releasing a new class of reinforcement learning algorithms, Proximal Policy Optimization (PPO), which perform comparably or better than state-of-the-art approaches while being much simpler to implement and tune.
We're open-sourcing OpenAI Baselines, our internal effort to reproduce reinforcement learning algorithms with performance on par with published results. We'll release the algorithms over upcoming months; today's release includes DQN and three of its variants.
We've created a robotics system, trained entirely in simulation and deployed on a physical robot, which can learn a new task after seeing it done once.
We’ve developed an unsupervised system which learns an excellent representation of sentiment, despite being trained only to predict the next character in the text of Amazon reviews.
We've discovered that evolution strategies (ES), an optimization technique that's been known for decades, rivals the performance of standard reinforcement learning (RL) techniques on modern RL benchmarks, while overcoming many of RL's inconveniences.
We're releasing Universe, a software platform for measuring and training an AI's general intelligence across the world's supply of games, websites and other applications.
Deep learning is an empirical science, and the quality of a group's infrastructure is a multiplier on progress. Fortunately, today's open-source ecosystem makes it possible for anyone to build great deep learning infrastructure.
OpenAI’s mission is to build safe AI, and ensure AI's benefits are as widely and evenly distributed as possible. We’re trying to build AI as part of a larger community, and we want to share our plans and capabilities along the way.
This post describes four projects that share a common theme of enhancing or using generative models, a branch of unsupervised learning techniques in machine learning.