Call for Abstracts
The Science Understanding through Data Science (SUDS) Conference Committee invites authors to submit an abstract for the 1st Annual SUDS Conference, held in-person at the California Institute of Technology from August 21st to 23rd, 2024. The complete schedule will be finalized after acceptance notifications.
Second Round Submissions are Open: https://cmt3.research.microsoft.com/SUDS2024/
Reviews and Timeline
- First Round Deadline:
May 31st, 2024 - Second Round Deadline: June 28th, 2024
- Conference: August 21st to 23rd, 2024
Abstracts will be reviewed for quality and alignment to one of the following sessions. Reviewers may request revisions from the author for clarity.
- Poster only
- Poster and Lightning Oral (10 min talk, 5 min Q&A)
- Contributed Oral (20 min talk, 10 min Q&A)
Due to the small scale of this inaugural year, we anticipate a limited number (<200) of attendees.
Introduction
SUDS began as an internal initiative at the Caltech/NASA Jet Propulsion Laboratory to develop and implement a roadmap for creating scientific understanding based on state-of-the-art data science methodologies. However, it is a need shared across organizations, institutions, research domains, and countries. The central organizing principle of this conference is to create the network of those working on SUDS under different names, research domains, and institutions.
The SUDS mission statement is “To increase the speed, depth, and rigor of scientific return by revealing new connections through data science”. The goal is to produce more than “an accurate prediction model,” but to facilitate new insight and scientific understanding.
We define data science to include innovative, data-driven approaches such as machine learning and artificial intelligence, often in the presence of existing physical models and/or with a need for uncertainty quantification.
Topics and Themes
For our inaugural meeting we are emphasizing, but not limiting to, the following science areas: Earth, Planetary, Heliophysics, Space Physics, Astronomy, and Cosmology, especially when incorporating remote sensing observations.
We are seeking abstracts that are aligned with one or more of the following categories:
- Collaborative works between physical scientists and data scientists
- Novel data science approaches motivated by physical science applications
- With emphasis on explainability, transparency, causality, or quantifiable uncertainty
- Novel physical science investigations enabled by data science approaches
- Operational data science-based implementations in science data pipelines
- Quantifying and assessing the effectiveness of physical science-data science collaboration
- Institutional strategies and organization of SUDS-like efforts
- Visions for the future of physical science-data science collaboration
We are not seeking the following:
- Data Science performance metric improvements on benchmark tasks (e.g. ImageNet, PDEBench, etc.)
- Simple automation of repetitive tasks
- Data science models trained on science data that do not connect back to science interpretation, insight, or benefit (a simple accuracy value is not sufficient)
- Processing pipelines, workflows, database management, scalable/cloud computing approaches
- Broad surveys or literature reviews
Authors that are unsure whether a submission would be appropriate for this conference may reach out to the organizing committee at sudsconf@jpl.nasa.gov
Format
The abstract has a limit of 500 words and one figure. The content of the abstract should be no longer than 1 page and clearly speak to both the data science approach and the physical science use case that was advanced. References may extend to a second page, and are not included in the abstract word count.
Abstract templates are available below:
Proceedings
Authors will have the opportunity to submit their abstract, oral presentation slides, and/or their posters to be shared as part of the conference proceedings on Zenodo. These submissions will be assigned a DOI for future reference.
Author Guidelines
Authors are encouraged to submit works published within the last year or work currently under review at other venues, as long as this is declared during submission. Authors should take care to abide by dual submission guidelines at other conferences/journals.
Plagiarism is prohibited by the SUDS Code of Conduct.
LLMs may be used by authors as a writing tool, as long as the author(s) take full responsibility for any and all content in their submission, including any misrepresentations, factual inaccuracies, and plagiarism.
The presenting author must register for and participate in the conference in order to have their abstracts published.