New Release
PixelCNN++: Improving the PixelCNN with discretized logistic mixture likelihood and other modifications
OpenAI unveils PixelCNN++, a model that enhances image generation with advanced likelihood modeling techniques.
OpenAI NewsOriginal →
2,476 stories
OpenAI unveils PixelCNN++, a model that enhances image generation with advanced likelihood modeling techniques.
Poorly defined reward functions in reinforcement learning can lead to unexpected and harmful AI behaviors.
OpenAI launches Universe, a groundbreaking platform for measuring AI's general intelligence across various applications.
OpenAI partners with Microsoft to enhance AI model testing through Azure's robust cloud infrastructure.
New research unveils effective count-based exploration techniques that enhance deep reinforcement learning performance.
OpenAI unveils new metrics to enhance the evaluation of decoder-based generative models.
New research uncovers links between GANs, inverse reinforcement learning, and energy-based models, promising advancements in AI training techniques.
OpenAI unveils RL², a groundbreaking approach to accelerate reinforcement learning performance.
OpenAI unveils a new variational lossy autoencoder designed to optimize data compression across various formats.
The neural GPU shows promise in algorithmic tasks but faces challenges with complex data structures and generalization.
OpenAI unveils a new method for enhancing deep learning using private training data, boosting model performance significantly.
Over 150 AI practitioners convened to explore collaborative machine learning at a groundbreaking self-organizing conference.
OpenAI unveils a groundbreaking model that enhances robotic performance by bridging simulation and real-world applications.
Open-source tools are revolutionizing deep learning infrastructure, enabling faster and more efficient model development.
The Machine Learning Unconference is set to provide attendees with essential updates and resources ahead of the event.
OpenAI announces a team expansion aimed at enhancing capabilities and driving innovation forward.
OpenAI emphasizes the importance of tackling significant scientific challenges for impactful AI advancements.
A new paper reveals urgent AI safety issues that require immediate attention from researchers and developers alike.
OpenAI outlines its technical goals focused on safety and equitable AI benefits for all.
OpenAI explores the future of generative models and their critical role in advancing AI capabilities.
New adversarial training methods enhance semi-supervised text classification, improving model robustness and accuracy.
OpenAI expands its team to accelerate AI research and innovation through new hires.
OpenAI Gym enters public beta, providing developers with powerful tools for reinforcement learning algorithm development.