New Frontier
"New Frontier" research encompasses the application of advanced machine learning techniques, particularly deep learning models like transformers and diffusion models, to diverse and challenging problems across various scientific domains. Current efforts focus on improving model efficiency and accuracy for tasks such as medical image segmentation, autonomous navigation and exploration, and large-scale data analysis, often leveraging techniques like attention mechanisms and novel fusion strategies for multimodal data. This research is significant because it pushes the boundaries of what's computationally feasible, leading to advancements in fields ranging from healthcare and robotics to materials science and financial technology.
Papers
Sequence Length Scaling in Vision Transformers for Scientific Images on Frontier
Aristeidis Tsaris, Chengming Zhang, Xiao Wang, Junqi Yin, Siyan Liu, Moetasim Ashfaq, Ming Fan, Jong Youl Choi, Mohamed Wahib, Dan Lu, Prasanna Balaprakash, Feiyi Wang
Pretraining Billion-scale Geospatial Foundational Models on Frontier
Aristeidis Tsaris, Philipe Ambrozio Dias, Abhishek Potnis, Junqi Yin, Feiyi Wang, Dalton Lunga