The CILVR Lab (Computational Intelligence, Learning, Vision, and Robotics) was founded in 2009 by Yann LeCun and Rob Fergus. Yann had joined NYU in 2003 after many years at AT&T, and Rob joined in 2008. CILVR is a group of faculty members, research scientists, postdocs, and students who work on AI, machine learning, and a wide variety of applications, notably computer perception, natural language understanding, robotics, and healthcare.
Work in Machine Learning spans multiple areas with a large emphasis on deep learning and flexible, interpretable, and scalable methods. Deep learning research includes self-supervised learning, representation learning, and generative models. More general work in machine learning includes signal processing, high-dimensional statistics, Gaussian Processes, and Kernel Learning. We also work on easy-to-use probabilistic inference and understanding the role of randomness and information in model building.
Other work in AI aims to understand the ingredients of intelligence. We use advances in machine intelligence to better understand human intelligence, and use insights from human intelligence to develop more fruitful kinds of machine intelligence. We have projects on making machine learning more natural and human-like, in order for AIs to continually learn, adapt, and reason in naturalistic environments.
Application areas of AI at CILVR include computer vision, healthcare and getting robots to generalize and adapt in the diverse world we live in.. Natural Language Processing is currently the largest, with projects on building high-quality training and evaluation data for Deep Learning-based models, applying these models to scientific questions in syntax and semantics, contributing to work on language model alignment and control, enabling reliable communication in natural language between machines and humans, and devising behavioral experiments and computational methods to study how people learn and understand language.
The CILVR Logo
The CILVR logo was devised in 2012 by Yann LeCun for the CILVR group’s original web site. The overall “Sigma” shape evokes the summation operation that is used to combine weights at a node in a typical neural network. The nodes and node connectors suggest the main building blocks in a neural network. Last, the “eyes” at the top are meant to convey the liveliness of the field.