XTRACTIS for Naval Security

Acoustic Detection of Underwater Mines

Benchmark vs. Logistic Regression, Random Forests, Boosted Trees & Neural Networks

Design an AI-based decision system that accurately and instantly detects underwater mines from sonar echoes to equip vessels, submarines, and drones with a detector making rational automated decisions.
Goals & benefits

Identify the frequency bands involved in the detection of underwater mines and enhance knowledge by helping submarine staff and acoustic experts understand the causal relationships between specific frequency bands, their combination, and the presence of a mine.

Help to design a virtual “Golden Ear” (expert in underwater acoustics) operating 24/7/365 with the same quality of decision, or to design by simulation undetectable mines.

Assist the military profession in making an earlier and more reliable decision, thanks to rapid, systematic and explainable detection process with limited sensors.

  • The top-model is a decision system composed of 23 unchained gradual rules aggregated into 2 disjonctive rules.
  • Each rule uses only some of the 29 variables that XTRACTIS identified as significant (out of the 60 Potential Predictors that are measures characterizing the energy in a specific frequency band).
  • Only a few rules are triggered at a time to compute the decision.

It has good Real Performance (on unknown data).

It computes real-time predictions up to 70,000 decisions/second, offline or online (API).

UC07 scores graph
LoR=Logistic Regression
RFo=Random Forests
BT=Boosted Trees
NN=Neural Networks

Detailed results and explanations in full document

Use Case 2024/02 (v4.0)

Powered by XTRACTIS® REVEAL 12.2.43406 (2022/09)


  1. Problem Definition
  2. XTRACTIS-induced Decision System
  3. XTRACTIS Process
  4. Top-Model Induction
  5. Explained Predictions for 3 unkown cases
  6. Top-Models Benchmark
  7. Appendices