Research
Researchers at NYU created the world's largest public dataset of AI-generated conversations to uncover how synthetic training data shapes the performance of large language models. Their work is helping make AI systems more capable, reliable and accessible while giving researchers across New York and beyond a powerful new resource for advancing open AI.
Cornell University researchers have used Empire AI to automate the workflow from electron microscopy images, helping to provide insights for technologies from advanced batteries and catalysts to semiconductors and next-generation manufacturing.
Researchers from the Icahn School of Medicine at Mount Sinai are developing a powerful AI system to help decode the genome and uncover how genetic variation influences health and disease.
Using Empire AI's advanced computing resources, a University of Buffalo team is building multimodal large language models (mLLM) capable of delivering real-time, interactive speech-language therapy that could create new opportunities for children to receive individualized support both in school and at home.
Scientists believe disruptions in the process of removing waste through the brain are associated with Alzheimer’s, stroke, hypertension, and traumatic brain injury. University of Rochester Prof. Douglas Kelley and his team are using Empire AI and advanced MRI imaging to reveal the hidden flow of waste-clearing fluids to reveal previously inaccessible aspects of brain function.
The rise of generative AI in music has introduced new opportunities but also significant concerns. A Cornell researcher is using Empire AI to develop AI tools that enhance rather than replace creativity, ensuring that artists remain at the center of the creative process.
Lung cancer is the leading cause of cancer-related death worldwide, but today’s screening programs face major barriers. Using Empire AI, an RPI researcher is developing a system that learns from large-scale, real-world data to identify subtle patterns that may not be clearly visible to clinicians to offer earlier, more accurate diagnoses.
Widely used large language models often function as “black boxes,” making their outputs difficult to predict or control and limiting their safe deployment in some sensitive domains, like healthcare, education and public services. A Cornell research team is working to allow users to better guide AI outputs.
