IEEE SMC-IT/SCC 2023: Explainable AI and Space Clouds

A mini workshop as a part of SMC-IT/SCC 2023, July 19, 2023, 1:00 - 4:45 pm, Lees-Kubota Lecture Hall, 133 Guggenheim, building 45 on the Caltech maps

Organizers: S. George Djorgovski (Caltech) and Richard Doyle (JPL/IEEE)

Image credit: Robert Hurt (Caltech/IPAC)

Motivation for the workshop:

As AI continues to grow in its capabilities and importance in all application domain, explainability of its results becomes a key issue, especially in the situations where AI is entrusted to make critical decisions. Explainable AI (XAI) is important both in the context of human understanding (data analytics) and trust (decision making). The term can be referring to an AI system’s ability to provide the rationale for actions and decisions, and how easily a human expert can interpret learned models. Both are relevant in the context of spacecraft autonomy.

Space exploration is enjoying early examples of shared services to ameliorate the burden of each individual mission having to embody all needed capabilities. Early cislunar planning is clearly embracing concepts for deploying common services and infrastructure. A recent study called Nebulae, sponsored by the Keck Institute of Space Studies, developed a vision for scalable deep-space computing, data and networking services. The goal of this workshop is to discuss how today’s thinking about space-based cloud services and edge computing can extend into the solar system, enabling new science and other mission concepts.

Workshop Structure:

The workshop will operate in two 90-minute tracks, each featuring three 20-minute invited talks, followed by a 25-minute Q&A and discussion involving panel of speakers, moderated by one of the workshop co-chairs. The first track will focus on Explainable AI, and the second track will focus on space-based cloud services. The workshop co-chairs will seek opportunities to cross-fertilize discussion across near-term and far-term considerations—with panel members, and with workshop attendees.

Speakers and Abstracts





1:30 - 1:35

Djorgovski & Doyle

Welcome and introductory comments

1:35 - 1:55

Chris Mattmann (JPL)

AI and Machine Learning from Back of the Napkin Sketch to Rovers on Mars

1:55 - 2:15

Sameer Singh (UCI)

Auditing Black Boxes: Model-Agnostic Explanations and Testing of Machine Learning

2:15 - 2:35

Ciro Donalek (Virtualitics)

Coupling Intelligent Exploration and Generative AI

2:35 - 3:00


Discussion: Explainable AI

3:00 - 3:15


3:15 - 3:35

Dan Crichton (JPL)

Considerations in Architecting Data-Driven Observing Systems for Space

3:35 - 3:55

Ashish Mahabal (Caltech)

Cloud Storage and Computing in Deep Space as Catalysts for Scientific Discovery

3:55 - 4:15

Laura Kennedy (MIT Lincoln Lab)

Space-based Clouds for Intelligence, Surveillance, and Reconnaissance Applications

4:15 - 4:40


Discussion: Space Clouds

4:40 - 4:45

Doyle & Djorgovski

Concluding remarks