Latest AI News

We’ve designed a method that encourages AIs to teach each other with examples that also make sense to humans. Our approach automatically selects the most informative examples to teach a concept—for instance, the best images to describe the concept of dogs—and experimentally we found our approach to be effective at teaching both AIs

We’ve co-authored a paper that forecasts how malicious actors could misuse AI technology, and potential ways we can prevent and mitigate these threats. This paper is the outcome of almost a year of sustained work with our colleagues at the Future of Humanity Institute, the Centre for the Study of Existential Risk, the Center for a New American Security, the Electronic Frontier Foundation, and others.

We’ve developed a simple meta-learning algorithm called Reptile which works by repeatedly sampling a task, performing stochastic gradient descent on it, and updating the initial parameters towards the final parameters learned on that task. Reptile is the application of the Shortest Descent algorithm to the meta-learning setting, and is mathematically similar to first-order MAML (which is a version of the well-known MAML algorithm) that only needs black-box

We’re releasing an experimental metalearning approach called Evolved Policy Gradients, a method that evolves the loss function of learning agents, which can enable fast training on novel tasks. Agents trained with EPG can succeed at basic tasks at test time that were outside their training regime, like learning to navigate to an object on a different side of the room from where it was placed during training.

We’re releasing an analysis showing that since 2012, the amount of compute used in the largest AI training runs has been increasing exponentially with a 3.4-month doubling time (by comparison, Moore’s Law had a 2-year doubling period)[^footnote-correction]. Since 2012, this metric has grown by more than 300,000x (a 2-year doubling period would yield only a 7x increase). Improvements in compute have been a key component of AI progress, so as long as this tr

We’re releasing the full version of Gym Retro, a platform for reinforcement learning research on games. This brings our publicly-released game count from around 70 Atari games and 30 Sega games to over 1,000 games across a variety of backing emulators. We’re also releasing the tool we use to add new games to the platform.

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. These results provide a convincing example that pairing supervised learning methods with unsupervised pre-training works very well; this is an idea that many have explored in the past, and we hope our result motivates f

We’ve trained an agent to achieve a high score of 74,500 on Montezuma’s Revenge from a single human demonstration, better than any previously published result. Our algorithm is simple: the agent plays a sequence of games starting from carefully chosen states from the demonstration, and learns from them by optimizing the game score using PPO, the same reinforcement learning algorithm that underpins OpenAI Five.

We introduce Glow, a reversible generative model which uses invertible 1x1 convolutions. It extends previous work on reversible generative models and simplifies the architecture. Our model can generate realistic high resolution images, supports efficient sampling, and discovers features that can be used to manipulate attributes of data. We’re releasing code for the model and an online visualization tool so people can explore and build on these results.

We’re proposing an AI safety technique called iterated amplification that lets us specify complicated behaviors and goals that are beyond human scale, by demonstrating how to decompose a task into simpler sub-tasks, rather than by providing labeled data or a reward function. Although this idea is in its very early stages and we have only completed experiments on simple toy algorithmic domains, we’ve decided to present it in its preliminary state because we

We’ve developed an energy-based model that can quickly learn to identify and generate instances of concepts, such as near, above, between, closest, and furthest, expressed as sets of 2d points. Our model learns these concepts after only five demonstrations. We also show cross-domain transfer: we use concepts learned in a 2d particle environment to solve tasks on a 3-dimensional physics-based robot.

We’re releasing CoinRun, a training environment which provides a metric for an agent’s ability to transfer its experience to novel situations and has already helped clarify a longstanding puzzle in reinforcement learning. CoinRun strikes a desirable balance in complexity: the environment is simpler than traditional platformer games like Sonic the Hedgehog but still poses a worthy generalization challenge for state of the art algorithms.

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. Since complex tasks tend to have noisier gradients, increasingly large batch sizes are likely to become useful in the future, removing one potential limit to further growth of AI systems. More broadly, these results show that neural network training need not be considered a mysterious art, but can

We’ve written a paper arguing that long-term AI safety research needs social scientists to ensure AI alignment algorithms succeed when actual humans are involved. Properly aligning advanced AI systems with human values requires resolving many uncertainties related to the psychology of human rationality, emotion, and biases. The aim of this paper is to spark further collaboration between machine learning and social science researchers, and we plan to hire s

We’ve made progress towards stable and scalable training of energy-based models (EBMs) resulting in better sample quality and generalization ability than existing models. Generation in EBMs spends more compute to continually refine its answers and doing so can generate samples competitive with GANs at low temperatures, while also having mode coverage guarantees of likelihood-based models. We hope these findings stimulate further research into this promisin

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. Both OpenAI Five and DeepMind’s AlphaStar had previously beaten good pros privately but lost their live pro matches, making this also the first time an AI has beaten esports pros on livestream.

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. MuseNet was not explicitly programmed with our understanding of music, but instead discovered patterns of harmony, rhythm, and style by learning to predict the next token in hundreds of thousands of MIDI files. MuseNet uses the same general-purpose unsupervised technology a

We’ve written a policy research paper identifying four strategies that can be used today to improve the likelihood of long-term industry cooperation on safety norms in AI: communicating risks and benefits, technical collaboration, increased transparency, and incentivizing standards. Our analysis shows that industry cooperation on safety will be instrumental in ensuring that AI systems are safe and beneficial, but competitive pressures could lead to a colle

Microsoft is investing $1 billion in OpenAI to support us building artificial general intelligence (AGI) with widely distributed economic benefits. We’re partnering to develop a hardware and software platform within Microsoft Azure which will scale to AGI. We’ll jointly develop new Azure AI supercomputing technologies, and Microsoft will become our exclusive cloud provider—so we’ll be working hard together to further extend Microsoft Azure’s capabilities i

We’re releasing the 774 million parameter GPT-2 language model after the release of our small 124M model in February, staged release of our medium 355M model in May, and subsequent research with partners and the AI community into the model’s potential for misuse and societal benefit. We’re also releasing an open-source legal agreement to make it easier for organizations to initiate model-sharing partnerships with each other, and are publishing a technical