Free Information Center
Jurassic-1 is a large-scale autoregressive language model developed by AI21 Labs, designed to generate human-like text for a variety of natural language processing tasks. It is notable for its size, architecture, and availability as part of AI21 Labs’ suite of AI tools.
Artificial intelligence in the United States encompasses the development, research, and application of AI technologies across various sectors, supported by government initiatives, academic research, and private industry. The U.S. has been a leading force in AI innovation, shaping global trends and ethical considerations.
The UNICRI Centre for AI and Robotics is a specialized unit within the United Nations Interregional Crime and Justice Research Institute, dedicated to advancing the understanding and responsible use of artificial intelligence and robotics to address crime and security challenges globally.
WaveGlow is a flow-based generative model designed for high-quality speech synthesis. It combines normalizing flows with a neural network architecture to generate audio waveforms directly from mel-spectrograms.
XLM-R (XLM-RoBERTa) is a multilingual transformer-based language model designed for cross-lingual understanding tasks. Developed by Facebook AI, it builds on the RoBERTa architecture and supports 100 languages.
VoxelNet is a deep learning architecture designed for three-dimensional object detection using LiDAR point cloud data. It integrates feature learning and 3D object detection into a single end-to-end trainable network, addressing challenges unique to 3D perception in autonomous systems.
Baidu Research is the research division of Baidu, a Chinese multinational technology company specializing in internet-related services and artificial intelligence. It focuses on advancing AI technologies through multiple institutes and collaborations worldwide.
MeshGraphNets are a class of machine learning models designed to simulate physical systems by leveraging graph neural networks on mesh-based data representations. They enable efficient and accurate prediction of complex physical phenomena by learning from simulation data structured as meshes.
A tool-augmented language model is a type of artificial intelligence system that combines traditional language understanding and generation capabilities with the ability to use external tools or knowledge sources to enhance performance. These models extend the functionality of standard language models by integrating APIs, databases, or other computational tools to provide more accurate, context-aware, and actionable outputs.
Recurrent neural networks (RNNs) are a class of artificial neural networks designed for processing sequential data, enabling them to maintain a memory of previous inputs.