Allen School researchers are at the forefront of exciting developments in AI spanning machine learning, computer vision, natural language processing, robotics and more.
We cultivate a deeper understanding of the science and potential impact of rapidly evolving technologies, such as large language models and generative AI, while developing practical tools for their ethical and responsible application in a variety of domains — from biomedical research and disaster response, to autonomous vehicles and urban planning.
Groups & Labs
H2 Lab
The H2 Lab addresses foundational problems in Artificial Intelligence and Natural Language Processing to develop general-purpose AI algorithms that represent, comprehend, and reason about diverse forms of data at large scale.
Bespoke Silicon Group
The Bespoke Silicon Group aims to bring hardware design to its highest art and rapidly conceive of, design and implement entirely new kinds of hardware faster than has ever been done before.
Faculty Members
Centers & Initiatives
The interdisciplinary DUB group at the University of Washington advances research, collaboration and teaching related to the interaction between design, people, and technology.
RAISE envisions a future where AI systems are developed and used in alignment with human ethics and values. With researchers from over a dozen labs across disciplines, RAISE is a leading center for research and education: building, evaluating, and envisioning AI technologies in the area of Responsible AI.
Highlights
Allen School News
The Institute of Electrical and Electronics Engineers (IEEE) recognized Kemelmacher-Shlizerman for her “contributions to face, body, and clothing modeling from large image collections,” including pioneering virtual try-on tools and bringing the technology to the mainstream.
Allen School News
A team of Allen School and Ai2 researchers were recognized for developing an efficient, scalable system for indexing petabyte-level text corpora with minimal storage overhead to better understand the data on which large language models are trained.
Allen School News
Allen School researchers led the development of a benchmark dataset of 26,000 real-world, open-ended queries to evaluate the creative generation of large language models. They discovered major LLMs all generate similar outputs as if they’re part of an Artificial Hivemind.