Embedded Files
ELLIS Robust LLMs Workshop
Keble College, Oxford, UK
ELLIS Workshop on Robustness in Large Language Models (RobustLLMs) is a two-day event hosted by the Oxford Department of Statistics at Keble College, Oxford and co-organised by ELISE. The workshop will feature keynotes and invited talks, discussions and poster sessions focused on the role of robustness in improving factuality and reasoning, defending against adversarial inputs, enhancing reliability for real-world applications, and dealing with hallucinations in LLMs.
Registrations are now closed, however you can attend remotely on zoom-Day1 and zoom- Day2.
Funding to cover travel and attendance costs is available for Europe-based students and early career researchers.
Update (11 June): Funding decisions have now been announced. If you have applied but not heard back, please contact us.
Confirmed Speakers
Alexander (Sasha) Rush (Cornell Tech, HuggingFace)
Beyza Ermiş (Cohere for AI)
Daniel Johnson (University of Toronto, Google DeepMind)
Frank Hutter (University of Freiburg)
Iryna Gurevych (TU Darmstadt)
Jonas Geiping (MPI Tübingen)
Roberta Raileanu (Meta AI)
Serena Booth (Brown University)
Sharon Li (UW Madison)
Subbarao Kambhampati (Arizona State)
Tim G. J. Rudner (New York University)
Yarin Gal (University of Oxford)
Call for Posters
We encourage you to share your research and contribute to a vibrant exchange of ideas by presenting a poster.
Submit your contribution via OpenReview by 31st May
Please submit a title, abstract, and a PDF of your contribution (could be a poster or paper). Accepted submissions will have their titles and abstracts published on the workshop website. We encourage submissions of already published work, work in progress, work under submission, or late-breaking results.
Update: Accepted posters have now been announced.
Schedule
Schedule
Day 1, 29 July
Day 1, 29 July
08:45-09:25 Walk in, Registration and Coffee
09:30-09:45 Opening Remarks
09:45-10:30 Talk 1: Iryna Gurevych
10:30-11:00 Talk 2: Frank Hutter (virtual)
11:00-11:30 Coffee Break
12:00-12:45 Talk 3: Sasha Rush
13:00-14:00 Lunch (served at Keble College)
14:00-14:45 Talk 5: Serena Booth
14:45-15:30 Talk 6: Subbarao Kambhampati
15:30-16:00 Talk 7: Sharon Li (virtual)
16:00-17:00 Poster Session & Coffee
17:00-17:45 Talk 4: Jonas Geiping
17:45-18:45 Walk in or around Oxford / Free time
19:00-19:30 Drinks Reception
19:30-21:00 Dinner at Keble College
Day 2, 30 July
Day 2, 30 July
08:45-09:00 Walk in, registration
09:00-09:45 Talk 8: Beyza Ermiş
09:45-10:30 Talk 9: Roberta Raileanu
10:30-11:00 Coffee Break
11:00-11:30 Talk 10: Tim G. J. Rudner (virtual)
11:30-12:15 Talk 11: Daniel Johnson
12:15-13:00 Talk 12: Yarin Gal
13:00-14:00 Lunch (served at Keble College)
14:15-15:15 Q&A session
15:15-15:30 Close
For more details, check out Schedule and Talks
Topics
Topics
Implications of Robustness on Safety, Hallucinations, Factuality and Reasoning:
Investigating robustness against adversarial inputs (prompt injections) specifically tailored to deceive LLMs.
Addressing robustness in the context of distributional shifts and their impact on LLM performance.
The role of robustness in improving factuality and reasoning of LLMs, detecting and mitigating hallucinations etc.
Enhancing LLM Reliability for Real-world Applications:
Techniques for uncertainty quantification in LLMs to improve decision-making reliability.
The role of diverse and challenging datasets in assessing and enhancing the reliability of LLMs.
Verification of LLM properties to ensure reliability and trustworthiness in real-world applications.
Societal, Ethical, and Legal Considerations:
Legal aspects concerning the robustness of LLMs, including liabilities related to misinterpretations and erroneous outputs.
Strategies for ensuring that LLMs are developed with fairness and without bias, promoting ethical AI practices.
Examining the robustness requirements of LLMs in critical sectors such as legal, healthcare, and content moderation, where the stakes are particularly high.
Innovation and Future Directions:
Novel approaches for improving the robustness of LLMs through architecture innovations, training methodologies, and data augmentation techniques.
The potential of hybrid models that combine the strengths of LLMs with other AI techniques for enhanced robustness.
Anticipating future challenges and opportunities in the evolving landscape of LLM robustness and reliability.
Venue
Venue
Oxford OX1 3PG,
United Kingdom
Contact
Contact
If you have any questions please contact ellis DOT robustmlworkshop AT gmail DOT com
Page updated
Google Sites
Report abuse