ICML 2024 Workshop on

Foundation Models in the Wild


July 26, 2024

The workshop will be held in a hybrid format.




News

  • Accepted papers are now available at OpenReview: https://openreview.net/group?id=ICML.cc/2024/Workshop/FM-Wild#tab-accept-poster. There are 92 poster papers and 3 oral papers!
  • Big thanks to Foundry, an incredible AI/computing startup that aims to orchestrate the world's compute, for generously providing $10,000 in Foundry credits distributed as four prizes for this workshop!
  • We will announce three prizes for outstanding papers, and one for an outstanding reviewer!

Overview

In the era of AI-driven transformations, foundation models (FMs), like large-scale language and vision models, have become pivotal in various applications, from natural language processing to computer vision. These models, with their immense capabilities, reshape the future of scientific research and the broader human society, but also introduce challenges in their in-the-wild deployments. The Workshop on FMs in the wild delves into the urgent need for these models to be useful when applied to our societies. The significance of this topic cannot be overstated, as the real-world implications of these models impact everything from daily information access to critical decision-making in fields like medicine and finance. Stakeholders, from developers to end-users, care deeply about this because the successful integration of FMs into in-the-wild frameworks necessitates a careful consideration of adaptivity, reliability and efficiency. Some of the fundamental questions that this workshop aims to address are:


  • Real-world Adaptation: In practical applications, how can we leverage the comprehensive knowledge in FMs to adapt them for specific domains, such as drug discovery, education, or clinical health?
  • Reliability and Responsibility: How can foundation models work reliably outside their training distribution? And how can we address issues like hallucination and privacy?
  • Safety, Ethics, and Fairness in Society: How do we ensure that the deployment of FMs preserving safety, ethics, and fairness within society, safeguarding against biases and unethical use?
  • Practical Limitations in Deployment: How can FMs tackle challenges in practical applications, such as system constraints, computational costs, data acquisition barriers, response time demands?


Call for Papers

We invite submissions from researchers in the fields of machine learning pertaining to foundation models and its in-the wild applications. Additionally, we welcome contributions from scholars in the natural sciences (such as physics, chemistry, and biology) and social sciences (including pedagogy and sociology) that necessitate the use of foundation models. In summary, our topics of interest include, but are not limited to:

  • Theoretical foundations of FMs in the wild
  • Empirical investigations into the in-the-wild deployments of various FMs
  • In-depth discussions exploring new applications of FMs in human society
  • Interventions during pre-training to enhance the downstream performance of FMs
  • Innovations in fine-tuning processes to bolster the adaptation of FMs to particular domains
  • Discussions on aligning models with potentially superhuman capabilities to human values
  • Advancements in the efficient fine-tuning and deployment of FMs to specific applications
  • Benchmark methodologies for assessing FMs in real-world settings
  • Issues of adaptivity, reliability and efficiency of FMs in broad applications

Submission URL:   https://openreview.net/group?id=ICML.cc/2024/Workshop/FM-Wild

Format:  All submissions must be in PDF format. Submissions are limited to four content pages, including all figures and tables; unlimited additional pages containing references and supplementary materials are allowed. Reviewers may choose to read the supplementary materials but will not be required to. Camera-ready versions may go up to five content pages.

Style file:   You must format your submission using the ICML 2024 LaTeX style file. For your convenience, we have modified the main conference style file to refer to the FM-WILD workshop: icml_fmwild.sty. Please include the references and supplementary materials in the same PDF as the main paper. The maximum file size for submissions is 50MB. Submissions that violate the ICML style (e.g., by decreasing margins or font sizes) or page limits may be rejected without further review.

Dual-submission policy:  We welcome ongoing and unpublished work. We will also accept papers that are under review at the time of submission, or that have been recently accepted without published proceedings.

Non-archival:  The workshop is a non-archival venue and will not have official proceedings. Workshop submissions can be subsequently or concurrently submitted to other venues.

Visibility:  Submissions and reviews will not be public. Only accepted papers will be made public.

Double-blind reviewing:   All submissions must be anonymized and may not contain any identifying information that may violate the double-blind reviewing policy. This policy applies to any supplementary or linked material as well, including code. If you are including links to any external material, it is your responsibility to guarantee anonymous browsing. Please do not include acknowledgements at submission time. If you need to cite one of your own papers, you should do so with adequate anonymization to preserve double-blind reviewing. Any papers found to be violating this policy will be rejected.

For any questions, please contact us at fmwild2024@googlegroups.com.


Important Dates

 

    Submission deadline: June 7, 2024, AOE (EXTENDED)

    Notification to authors: July 3, 2024, AOE (EXTENDED)

    Camera-ready deadline: July 17, 2024, AOE (EXTENDED)

    Workshop Data: July 26, 2024, AOE

Schedule

This is the tentative schedule of the workshop. All slots are provided in Central European Summer Time (CEST).

Morning Session


08:50 - 09:00 Introduction and opening remarks
09:00 - 09:30 Invited Talk 1: Dakuo Wang
09:30 - 10:00 Invited Talk 2: David Alvarez-Melis
10:00 - 10:15 Oral Presentation 1: Parameter-Efficient Quantized MoE Meets Vision-Language Instruction Tuning for Semiconductor Electron Micrograph Analysis
10:15 - 11:15 Poster Session 1
11:15 - 11:45 Invited Talk 3: Boran Han
11:45 - 12:00 Oral Presentation 2: RouteFinder: Towards Foundation Models for Vehicle Routing Problems
12:00 - 12:30 Invited Talk 4: Hannah Kerner
12:30 - 13:30 Break

Afternoon Session


13:30 - 13:45 UK AI Safety Institute: Empirically Assessing AI's Risks & Advancing Systemic Safety
13:45 - 14:00 Oral Presentation 3: DigiRL: Training In-The-Wild Device-Control Agents with Autonomous Reinforcement Learning
14:00 - 14:30 Invited Talk 5: Steven Wu
14:30 - 15:30 Poster Session 2
15:30 - 16:00 Invited Talk 6: Pang Wei Koh
16:00 - 16:30 Invited Talk 7: Jimeng Sun
16:30 - 17:00 Invited Talk 8: Sheng Wang
 

Invited Speakers




Dakuo Wang

Northeastern University

David Alvarez-Melis

Harvard University

Boran Han

AWS AI Research

Pang Wei Koh

University of Washington

Jimeng Sun

University of Illinois Urbana-Champaign



Sheng Wang

University of Washington

Hannah Kerner

Arizona State University

Steven Wu

Carnegie Mellon University

Workshop Organizers




Xinyu Yang

Carnegie Mellon University

Bilge Acun

Meta AI

Kamalika Chaudhuri

University of California San Diego

Beidi Chen

Carnegie Mellon University & Meta AI

Giulia Fanti

Carnegie Mellon University



Junlin Han

University of Oxford & Meta AI

Shengbang Tong

New York University

Philip Torr

University of Oxford

Hao Wang

Rutgers University

Cathy Wu

Massachusetts Institute of Technology



Huaxiu Yao

UNC-Chapel Hill

James Zou

Stanford University