Enabling Onboard Detection of Events of Scientific Interest for the Europa Clipper Spacecraft

Wagstaff, Kiri L.; Doran, Gary; Davies, A. G.; Anwar, Saadat; Chakraborty, Srija; Cameron, M. E.; Daubar, I. J.; Phillips, C. B.

doi:10.1145/3292500.3330656

Cited by 17 publications

(13 citation statements)

References 16 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Generative adversarial networks (GANs) (Goodfellow et al 2014), which have been successfully used for learning data-generating distributions for complex datasets (e.g., Antipov et al 2017;Dong et al 2018), have been recently proposed for novelty detection (Schlegl et al 2017;Akcay et al 2018;Zenati et al 2018b). The Reed-Xiaoli (RX) method, which computes pixel-wise anomaly scores using the Mahalanobis distance between the pixel and a background distribution (Reed and Yu 1990), and its kernel variants are widely used for unsupervised anomaly detection in multispectral and hyperspectral images (e.g., Kwon and Nasrabadi 2005;Molero et al 2013;Zhou et al 2016;Ayhan et al 2017;Wagstaff et al 2019). Though RX is usually used for detecting global or local outliers/anomalies, it can be used for novelty detection by computing the background statistics from the typical training dataset as proposed in our study.…”

Section: Related Workmentioning

confidence: 99%

“…Similarly, the reconstruction error in PCA assesses novelty as the distance of the input data from the mean in the space of the low-variance principal components (Chang and Chiang 2002). The primary difference between PCA and RX is that the reconstruction error in PCA measures the residual information along the components k + 1 to n where k is number of high-variance components retained in the projection matrix and n is the number of input data features (n is the number of pixels for the flattened image representation), whereas the inverse covariance matrix in RX includes all components (Chang and Chiang 2002;Wagstaff et al 2019).…”

Section: Rx Detectormentioning

confidence: 99%

See 1 more Smart Citation

Comparison of novelty detection methods for multispectral images in rover-based planetary exploration missions

Kerner

Wagstaff

Bue

et al. 2020

Data Min Knowl Disc

Self Cite

View full text Add to dashboard Cite

Science teams for rover-based planetary exploration missions like the Mars Science Laboratory Curiosity rover have limited time for analyzing new data before making decisions about follow-up observations. There is a need for systems that can rapidly and intelligently extract information from planetary instrument datasets and focus attention on the most promising or novel observations. Several novelty detection methods have been explored in prior work for three-channel color images and non-image datasets, but few have considered multispectral or hyperspectral image datasets for the purpose of scientific discovery. We compared the performance of four novelty detection methods—Reed Xiaoli (RX) detectors, principal component analysis (PCA), autoencoders, and generative adversarial networks (GANs)—and the ability of each method to provide explanatory visualizations to help scientists understand and trust predictions made by the system. We show that pixel-wise RX and autoencoders trained with structural similarity (SSIM) loss can detect morphological novelties that are not detected by PCA, GANs, and mean squared error autoencoders, but that the latter methods are better suited for detecting spectral novelties—i.e., the best method for a given setting depends on the type of novelties that are sought. Additionally, we find that autoencoders provide the most useful explanatory visualizations for enabling users to understand and trust model detections, and that existing GAN approaches to novelty detection may be limited in this respect.

show abstract

Section: Related Workmentioning

confidence: 99%

Section: Rx Detectormentioning

confidence: 99%

Comparison of novelty detection methods for multispectral images in rover-based planetary exploration missions

Kerner

Wagstaff

Bue

et al. 2020

Data Min Knowl Disc

Self Cite

View full text Add to dashboard Cite

show abstract

“…Another advantage that less complex models can offer is their relatively modest hardware requirements. This becomes a key decision point in resource constrained environments, as shown by Wagstaff et al [31]. They worked on deploying ML models to a range of scientific instruments onboard Europa Clipper spacecraft.…”

Section: Model Selectionmentioning

confidence: 99%

“…Essentially, performance metrics should reflect audience priorities. For instance Sato et al [45] recommend validating models for bias and fairness, while in the case described by Wagstaff et al [31] controlling for consumption of spacecraft resources is crucial.…”

Section: Requirement Encodingmentioning

confidence: 99%

Challenges in Deploying Machine Learning: a Survey of Case Studies

Paleyes,

Urma,

Lawrence

2020

Preprint

View full text Add to dashboard Cite

In recent years, machine learning has received increased interest both as an academic research field and as a solution for real-world business problems. However, the deployment of machine learning models in production systems can present a number of issues and concerns. This survey reviews published reports of deploying machine learning solutions in a variety of use cases, industries and applications and extracts practical considerations corresponding to stages of the machine learning deployment workflow. Our survey shows that practitioners face challenges at each stage of the deployment. The goal of this paper is to layout a research agenda to explore approaches addressing these challenges.

show abstract

“…For instance, (Kerner et al, 2019) utilized a machine learning algorithm to detect novel geologic features in multispectral images of the Martian surface. (Wagstaff et al, 2019) used machine learning methods to study thermal anomalies, compositional anomalies, and plumes of icy matter from Europa's subsurface ocean. (Nguyen et al, 2019) utilized machine learning techniques to automatically detect the terrestrial bow shock and magnetopause from in-situ data, and (Lazzús et al, 2017) used machine learning algorithms to forecast the Dst index of the Earth.…”

Section: Introductionmentioning

confidence: 99%