New Release
Improving language understanding with unsupervised learning
OpenAI unveils a groundbreaking unsupervised learning system that sets new benchmarks in language understanding.
OpenAI NewsOriginal →
2,472 stories
OpenAI unveils a groundbreaking unsupervised learning system that sets new benchmarks in language understanding.
OpenAI's GamePad introduces a groundbreaking interactive learning environment for theorem proving.
OpenAI opens applications for its Fall 2018 Fellowship program, offering a unique opportunity for aspiring AI researchers.
OpenAI's Gym Retro expands its library, offering over 1,000 games for AI reinforcement learning research.
AI training compute usage has skyrocketed, doubling every 3.4 months and surpassing Moore's Law by a significant margin.
OpenAI introduces a novel AI safety technique that uses debate to enhance decision-making and reliability.
OpenAI's Evolved Policy Gradients transform AI training, enabling agents to adapt swiftly to new tasks and environments.
OpenAI introduces a new benchmark aimed at enhancing generalization in reinforcement learning models.
OpenAI launches a contest to push the boundaries of reinforcement learning and transfer learning techniques.
OpenAI unveils a new method to enhance policy gradient efficiency in reinforcement learning through action-dependent baselines.
OpenAI's inaugural hackathon brings together 100 AI enthusiasts to spark innovation and collaboration in AI development.
New research integrates optimal transport techniques into GANs, enhancing their performance and image quality.
First-order meta-learning algorithms promise reduced costs and improved adaptability for AI models in dynamic environments.
OpenAI unveils Reptile, a new meta-learning algorithm designed for efficient task adaptation in AI models.
OpenAI launches Scholars program to support underrepresented groups in deep learning with stipends and mentorship.
OpenAI unveils new algorithms to enhance exploration strategies in meta-reinforcement learning.
OpenAI introduces eight new simulated robotics environments to boost AI training for physical robots.
OpenAI launches a new initiative to advance multi-goal reinforcement learning in robotics, inviting collaboration among researchers.
OpenAI announces an exciting hackathon in San Francisco, set to inspire developers and AI enthusiasts alike.
A new paper from top researchers outlines risks and strategies to combat malicious uses of AI technology.
OpenAI announces new donors joining its mission to enhance AI research and development.
OpenAI unveils a groundbreaking method for AIs to teach each other using relatable human examples.
OpenAI unveils an AI system that automates entity disambiguation using 100 distinct types.