OpenAI Scholars 2020: Final projects
OpenAI's third class of Scholars presented their final research projects at a virtual event. The projects were developed over five months and showcase...
OpenAI's third class of Scholars presented their final research projects at a virtual event. The projects were developed over five months and showcase...
The Reformer is a new type of transformer-based model that aims to reduce memory usage and improve training speed. It achieves this by using locality-...
OpenAI is collaborating on two NeurIPS 2020 competitions with AIcrowd, Carnegie Mellon University, and DeepMind. The competitions will use the Procgen...
Researchers at OpenAI have found that large transformer models can generate coherent images when trained on pixel sequences. They compared the quality...
OpenAI has released an API that allows developers to access its latest AI models. The API will enable users to integrate these models into their appli...
Deep reinforcement learning agents often face a trade-off between exploiting known solutions and exploring new possibilities. If they commit too quick...
Researchers at OpenAI have found that large language models can learn new tasks with just a few examples, similar to humans. This ability is called 'f...
OpenAI's analysis shows that the amount of computing power needed to train a neural network to perform well on image classification tasks has been dec...
OpenAI has released Jukebox, a neural network that generates music, including basic singing, in various styles and genres. The model produces raw audi...
OpenAI has contributed to a report on improving verifiability in AI development. The report, written by 58 co-authors from 30 organizations, outlines ...
OpenAI has released a tool called OpenAI Microscope, which provides visualizations of significant layers and neurons in eight vision models. These mod...
The article discusses the Transformer family, a set of neural network architectures introduced in 2017 by Vaswani et al. Since its introduction, the T...