Machine learning can be used to significantly expand the capabilities of
remote agents operating in space missions. For example, spacecraft could
intelligently filter their observations to make the best use of available
bandwidth or rovers with learning capabilities could more thoroughly and
more quickly explore new environments. Autonomous robots can play a key
role in creating a successful human presence on the Moon and Mars, both
before humans arrive and in collaboration with them once humans are on site.
However, care must be exercised in applying and developing techniques which
will truly operate without human intervention. The risks and possible
safety implications need to be well understood.
The purpose of this special issue is to collect recent advances in
machine learning for remote space or planetary environments and to
identify novel space applications where machine learning could
significantly increase capabilities, robustness, and/or efficiency.
Key topics of interest include:
- How to perform machine learning in a high-risk, remote environment
- Learning with resource constraints (computation, memory, etc.)
- Multi-instrument machine learning
- Multi-mission machine learning
- Novel applications and uses of on-board machine learning
- On-board machine learning for computer vision and image analysis
- Prioritizing or subsampling data for downlink
- Active selection of new observational targets
- How to evaluate and validate machine learning methods prior to
deployment on-board a spacecraft
- Methods for safe real-time learning
- Methods that trade off exploration and exploitation, given mission
science goals and safety/reliability requirements
- Methods for reducing risk and increasing acceptance of machine
learning in space flight missions
- A survey of space-borne machine learning accomplishments
We encourage all prospective authors to email us with a brief summary
the paper concept for feedback, especially for surveys or papers
focused on applications.
Submissions are expected to represent high-quality, significant
contributions in the area of machine learning algorithms and/or
applications. Authors should follow standard formatting guidelines for
Machine Learning manuscripts.