Uploaded June 2022 | Updated September 2026, 15 hours ago
The success of large AI models in challenging (yet conceptually simple) perception and comprehension tasks (e.g., generating images from a text description) is motivating new applications of these methods in areas previously considered to require human expertise. For example, deep learning has shown promise in replacing physics-based models to predict the weather, simultaneously learning the physics and the parameters directly from observations, ignoring hundreds of years of research in fluid dynamics. We have been considering AI applications in small-scale, high-value, high-expertise settings, characterized by limited training data, complex multi-modal features, ambiguous and conflicting success metrics, and high risk/reward potential. With the commoditization of methods (deep neural networks trained by gradient descent), progress in these settings is driven almost entirely by the careful specification of the task, the quality of the training data, and the fidelity of the evaluation, all of which provide an opportunity for domain experts to engage more directly in the design and even implementation of AI systems -- curation as programming. I’ll describe some applications my group is pursuing including understanding how the use of visualization relates to impact in the scientific literature, recovering missing data in urban mobility, summarizing public opinion around societal goals, identifying speakers and agenda items in local council meetings, and extracting information from legal documents. I’ll describe some patterns we find, with some emphasis on how best to make use of expert-provided ontologies when they are available, and on some preliminary results in unlearning biases during fine-tuning. I’ll end with some discussion topics for directions going forward, focusing on the scientific literature as the application domain.
Bio:
https://faculty.washington.edu/billhowe/bio.html
Bill Howe is an Associate Professor in the Information School and Adjunct Associate Professor in the Allen School of Computer Science & Engineering and the Department of Electrical Engineering. His research interests are in data management, machine learning, and visualization, particularly as applied in the physical and social sciences. As Founding Associate Director of the UW eScience Institute, Dr. Howe played a leadership role in the Moore-Sloan Data Science Environment program through a $32.8 million grant awarded jointly to UW, NYU, and UC Berkeley, and founded UW’s Data Science for Social Good Program. With support from the MacArthur Foundation, NSF, and Microsoft, Howe directs UW’s participation in the Cascadia Urban Analytics Cooperative. He founded the UW Data Science Masters Degree, serving as its inaugural Program Chair, and created a first MOOC on data science that attracted over 200,000 students. His research has been featured in the Economist and Nature News, and he has authored award-winning papers in conferences across data management, machine learning, and visualization. He has a Ph.D. in Computer Science from Portland State University and a Bachelor’s degree in Industrial & Systems Engineering from Georgia Tech.
The success of large AI models in challenging (yet conceptually simple) perception and comprehension tasks (e.g., generating images from a text description) is motivating new applications of these methods in areas previously considered to require human expertise. For example, deep learning has shown promise in replacing physics-based models to predict the weather, simultaneously learning the physics and the parameters directly from observations, ignoring hundreds of years of research in fluid dynamics. We have been considering AI applications in small-scale, high-value, high-expertise settings, characterized by limited training data, complex multi-modal features, ambiguous and conflicting success metrics, and high risk/reward potential. With the commoditization of methods (deep neural networks trained by gradient descent), progress in these settings is driven almost entirely by the careful specification of the task, the quality of the training data, and the fidelity of the evaluation, all of which provide an opportunity for domain experts to engage more directly in the design and even implementation of AI systems -- curation as programming. I’ll describe some applications my group is pursuing including understanding how the use of visualization relates to impact in the scientific literature, recovering missing data in urban mobility, summarizing public opinion around societal goals, identifying speakers and agenda items in local council meetings, and extracting information from legal documents. I’ll describe some patterns we find, with some emphasis on how best to make use of expert-provided ontologies when they are available, and on some preliminary results in unlearning biases during fine-tuning. I’ll end with some discussion topics for directions going forward, focusing on the scientific literature as the application domain.
Bio:
https://faculty.washington.edu/billhowe/bio.html
Bill Howe is an Associate Professor in the Information School and Adjunct Associate Professor in the Allen School of Computer Science & Engineering and the Department of Electrical Engineering. His research interests are in data management, machine learning, and visualization, particularly as applied in the physical and social sciences. As Founding Associate Director of the UW eScience Institute, Dr. Howe played a leadership role in the Moore-Sloan Data Science Environment program through a $32.8 million grant awarded jointly to UW, NYU, and UC Berkeley, and founded UW’s Data Science for Social Good Program. With support from the MacArthur Foundation, NSF, and Microsoft, Howe directs UW’s participation in the Cascadia Urban Analytics Cooperative. He founded the UW Data Science Masters Degree, serving as its inaugural Program Chair, and created a first MOOC on data science that attracted over 200,000 students. His research has been featured in the Economist and Nature News, and he has authored award-winning papers in conferences across data management, machine learning, and visualization. He has a Ph.D. in Computer Science from Portland State University and a Bachelor’s degree in Industrial & Systems Engineering from Georgia Tech.
