Cornell Tech faculty made a strong showing at the 2025 Conference on Neural Information Processing Systems, presenting 23 research papers at the event held Dec. 2-7 in San Diego. NeurIPS draws thousands of scholars and industry leaders each year and is widely recognized as a leading forum for breakthroughs in Artificial Intelligence, computational neuroscience, statistics, and large-scale modeling. The Cornell Chronicle reported the roundup on December 9, 2025.
The Cornell Tech contributions span multiple technical fronts. Teams highlighted work to safeguard data privacy and to strengthen evaluation standards for Artificial Intelligence systems, while other papers focused on boosting the speed and efficiency of large language models. Additional projects unveiled tools for analyzing environmental and health interventions, methods for matching images to architectural plans, and approaches for generating realistic 3D scenes with unprecedented efficiency.
Researchers framed these advances as having broad practical implications. Improved privacy and evaluation practices aim to make deployments more trustworthy; performance gains in large language models target efficiency in real-world applications. The environmental and health analysis tools have potential impact for public health decision making, the image-to-plan work informs robotics and urban design workflows, and the efficient 3D generation methods promise richer immersive media and simulation capabilities. Together, the 23 papers illustrate Cornell Tech’s push to advance Artificial Intelligence research across theory, tools, and applied domains.
