💻 Technology: developing new algorithms and metrics that
promote transparency, robustness, safety, and fairness, with goals inspired by issues unearthed by the 🖍️ Design Trail (led by UMD CS's Tom Goldstein)
>
Hal Daumé III
@hal@sigmoid.social
- 2.8K Followers
- 87 Following
- 105 Posts
Natural language processing, machine learning, AI prof at UMD CS, member of CLIP & HCIL; Researcher Microsoft Research NYC. Views my own. Fun: 🧗🧑🍳🧘⛷️🏕️ he/him. Was @haldaume.
All of this was made possible by an amazing leadership team -- Susan A. Aaronson, David Broniatowski, Virginia Byrne, Tom Goldstein, and Katie Shilton -- as well as the invaluable of UMD's VPR's Amanda Dykema, amazing staff support at our institutions, and funding from the NSF.
Many thanks to the friends and colleagues who helped us by providing "red team" feedback on both the proposal and our mock reverse site visit.
And of course to all researchers involved: trails.umd.edu/faculty
If you're a researcher who would like to get involved, we're working on mechanisms for that, check in the Fall.
If you're at a public/civil society organization and would like to connect, email us at trails@umiacs.umd.edu.
Media inquiries to Tom Ventsias: tventsia@umd.edu
>
We have a few upcoming events, including:
A public Generative AI 101 Webinar: https://trails.umd.edu/events/generativeai101webinar
A chat with Congressman Don Beyer about AI and governance: https://trails.umd.edu/events/donbeyer
Details: http://trails.umd.edu/events
>
Generative AI 101 Webinar — NSF Institute for Trustworthy AI in Law & Society (TRAILS)
Generative AI is many things: it is simultaneously scary, useful, flexible, opaque, exciting, and dangerous. Researchers, policymakers, creatives, and executives around the world are turning to these models to answer complex questions, rethink practices and procedures, and save time. Large language
If you're an undergraduate or graduate student at UMD, GW, or Morgan State, we will also shortly announce student fellowships and seed grants.
Anyone can follow trails_ai (on Twitter) or join our announcements mailing list (https://lists.umiacs.umd.edu/mailman/listinfo/trails-announce) to stay informed!
>
If this sounds like something you'd like to join, we have several openings (now, or very soon):
- Postdoctoral Fellowships at
UMD and GW.
- Managing Director position at UMD
- Outreach Coordinator at GW
Details: https://trails.umd.edu/getinvolved
>
This effort unifies:
UMD's expertise in ML, AI, & understanding of sociotechnical systems...
GW's expertise in law, policy, governance, HCI, and systems engineering...
Morgan State's expertise in deep community-based research and forward-looking education.
>
⚖️ Governance: analyzing the governance of AI and its inputs such as data at the national and international levels (led by GW's Susan A. Aaronson)
>
🧠 Sense-Making: understanding how people respond to AI through the development of psychometric measures and experimental paradigms to assess system transparency (led by GW's site lead David Broniatowski with expertise from Cornell DecisionsLab's Valerie Reyna)
>
Our research is divided into four "trails":
🖍️ Design: developing participatory methods specific to AI, to ensure that AI reflects diverse stakeholder values (led by
UMD iSchool's Katie Shilton with
Morgan State's Virginia Byrne leading efforts for education)
>