Regular papers
Research papers of up to 6 pages, excluding references and supplementary material.
AdvML Frontiers aims to form an interdisciplinary community of human-centered researchers within trustworthy AI. Although individual researchers rely on human-centered methods and several published SaTML papers include human-factors components, there is not yet a dedicated community around this work. The workshop welcomes senior researchers and PhD students already working on—or interested in—human factors in trustworthy machine learning, with a low barrier to entry for those who want to learn, participate, and network.
Important dates
- Submission deadline
- Decision notification
- Camera-ready deadline
- Workshop
Call for Papers
We invite human-centered work on trustworthy machine learning systems, especially interdisciplinary research spanning machine learning, security, safety, HCI, ethics, and the social sciences.
Topics of interest
- Impact of Insecurity/Security ML on People: how trustworthy or untrustworthy ML can benefit, harm, or otherwise affect people.
- Needs and Acceptability of Trustworthy ML: what forms of trustworthy ML are needed in particular sociotechnical circumstances and how needs vary across people, contexts, and experience levels.
- Empirical Methods for Measuring Effectiveness In Practice: human-subjects experiments, internet measurement, surveys, interviews, ethnography, case studies, observational studies, and related social-science methods.
- Usable SaTML Systems: studies that evaluate or improve the interfaces and designs of systems relevant to secure and trustworthy machine learning.
- Human-Centric SaTML-Relevant Processes: evaluations of human-in-the-loop processes including red teaming, safety practices, and human considerations around ML development and use.
Blue Sky, Position, or Demo papers
Up to 4 pages on visionary ideas, long-term challenges, current debates, overlooked questions, or demonstrations.
Non-archival and open to work in progress. Submissions do not need to be anonymized. Concurrent and previously archived work is welcome, and accepted papers will receive either a contributed spotlight or poster presentation.
The submission site will be announced in the near future.
Keynote Speakers
To be announced.
Program
May 3, 2027 · 9 AM–4 PM · Reykjavík, Iceland · Venue forthcoming
Welcome and introductions
15 minutes
Keynote 1
45 minutes · Speaker and talk to be announced
Details
Speaker, title, and abstract to be announced.
Paper Talks 1
30 minutes
Details
Paper titles and presentation order will be added after acceptance decisions.
Coffee break and poster set-up
15 minutes
Keynote 2
45 minutes · Speaker and talk to be announced
Details
Speaker, title, and abstract to be announced.
Paper Talks 2
30 minutes
Details
Paper titles and presentation order will be added after acceptance decisions.
Lunch
60 minutes
Poster session
90 minutes
Speed Mentoring
1:00–1:30 PM
Keynote 3
45 minutes · Speaker and talk to be announced
Details
Speaker, title, and abstract to be announced.
Panel discussion
45 minutes
Details
Discussion with the keynote speakers and selected workshop organizers. The final panel lineup and topic will be announced with the program.
Closing remarks
5 minutes
Organizers
Pin-Yu Chen
IBM Research
Kathrin Grosse
Independent Researcher
Sanmi Koyejo
Stanford University
Sijia Liu
Michigan State University
Jaron Mink
Arizona State University
Yao Qin
UC Santa Barbara
Lea Schönherr
CISPA
Eric Wong
University of Pennsylvania
Dongxiao Zhu
Wayne State University