Zoetrope genetic programming for regression

Boisbunon, Aurélie; Fanara, Carlo; Grenet, Ingrid; Daeden, Jonathan; Vighi, Alexis; Schoenauer, Marc

doi:10.1145/3449639.3459349

Cited by 4 publications

(3 citation statements)

References 35 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…1) A genetic programming and frugal model obtained by ZGP algorithm (Zoetrope Genetic Programming) [15] that segments and filters only sea related patches for further processing. 2) A neural network that projects patches into a more robust and compact feature space, which restitutes the highlevel semantics of the scene on the patch.…”

Section: A Architecture Descriptionmentioning

confidence: 99%

“…Those with a significant anomaly score are kept for the next step. 4) A ZGP model again [15] that precisely characterizes some types of anomaly when anomaly detections occur.…”

Section: A Architecture Descriptionmentioning

confidence: 99%

“…Its core training principle involves evolving individuals represented by mathematical expressions that combine input variables and constants. After undergoing mutation and recombination processes across several generations, the best individual is selected as the final model, as outlined in [15]. The resulting models consist of a simple mathematical formula that combines the model's input variables and numerical constants.…”

Section: B Architecture Justificationmentioning

confidence: 99%

See 2 more Smart Citations

Onboard Anomaly Detection for Marine Environmental Protection

Goudemant,

Francesconi,

Aubrun

et al. 2024

IEEE J. Sel. Top. Appl. Earth Observations Remote Sensing

Self Cite

View full text Add to dashboard Cite

This article describes a comprehensive artificial intelligence pipeline for detecting threatening events in the marine environment, which is intended to be run on-board Earth observation satellites and globally contribute to the preservation of the marine environment. We employed a self-supervised neural network-based anomaly detection technique to identify a wide range of potentially unknown events (pollutions). This method consist in identifying deviations from the learned "normal" water state as abnormal events. Then, we can better characterize some of the anomalies by incorporating other, more selective models into our processing pipeline to prioritize targeted actions over specific anomalies (e.g., oil spills). This approach's effectiveness and versatility are demonstrated through its selection in two European Space Agency competitive challenges aimed at integrating artificial intelligence into two Earth observation demonstration missions: Φsat-2 and IMAGIN-e. The article details the processing pipeline design, the generated datasets for training and testing steps of the two challenges, and the detection effectiveness and hardware efficiency (including throughput and latency) when embedded on missions-representative hardware. Our primary contribution is the proposal of an operational pipeline with the constraints to (1) detect multiple and unknown events, (2) require only "weakly" labelled data and (3) be computationally efficient enough to run on low-power targets on-board Earth observation payloads. These three assets make this pipeline easily adaptable to a wide range of missions.

show abstract

Section: A Architecture Descriptionmentioning

confidence: 99%

“…Those with a significant anomaly score are kept for the next step. 4) A ZGP model again [15] that precisely characterizes some types of anomaly when anomaly detections occur.…”

Section: A Architecture Descriptionmentioning

confidence: 99%

Section: B Architecture Justificationmentioning

confidence: 99%

See 1 more Smart Citation