Highly Intelligible

XTRACTIS-induced models are intelligible: humans can fully understand the internal decision-making logic of the AI systems automating the decisions.

This chart shows the INTELLIGIBILITY “#1 ranking” for 16 Public Use Cases.

No Data Found

Highly Predictive

XTRACTIS-induced models are robust: their performance is at least as high as mainstream AI and data-driven modeling techs, as benchmarks prove. 

This chart shows the PERFORMANCE “#1 ranking” for 16 Public Use Cases.

No Data Found

Public Use Cases on Real World Data

Our Use Cases present the results of XTRACTIS modeling in many application sectors and include complete Benchmarks against Neural Networks, Boosted Trees, Random Forests, and Logistic Regression.

These studies illustrate the ability of Xtractis to automatically induce knowledge in the form of predictive and intelligible mathematical relationships from real-world data (public data or authorized private data).

For each application, we also show how this knowledge is used to obtain explained predictions for situations unknown from the learning data set.

Choose the sector to access the corresponding Use Cases.

HEALTH / PHARMA
INDUSTRY / R&D
DEFENSE / CYBER / SECURITY
BUSINESS / FINANCE
  • Discovery of strategies to detect Parkinson's
  • from patients simple voice recording
  • to make an explainable and accurate automated medical diagnosis
  • Discovery of strategies to detect kidney disease
  • from the patient record and blood measures
  • to make an explainable and accurate automated medical diagnosis
HEALTH / PHARMA | Jan 2023

Genetic Identification of Lung Cancer

  • Discovery of strategies to identify 2 types of lung cancer
  • from the expression levels of 12,533 genes and for a small number of patients
  • to make an explainable and accurate automated medical diagnosis
HEALTH / PHARMA | Oct 2022

Genetic Diagnosis of Prostate Cancer

  • Discovery of strategies to detect cancer
  • from the expression levels of 12,600 genes and for a small number of patients
  • to make an explainable and accurate automated medical diagnosis
  • Discovery of strategies to identify heart conditions
  • from signal characteristics of fetal heart rate and uterine contractions
  • to make an explainable and accurate automated medical diagnosis
HEALTH / PHARMA | Sep 2022

Spectrometric Diagnosis of Ovarian Cancer

  • Discovery of strategies to detect cancer
  • from the 15,154 mass/charge ratios originating from the protein spectrum of serum samples and for a small number of patients
  • to make an explainable and accurate automated medical diagnosis
  • Discovery of strategies to detect cancer
  • from topological characteristics of mammary cells
  • to make an explainable and accurate automated medical diagnosis
Edit Template
  • Automatic exploration of the 86-dimension continuous decision space of possible molecules formed by the 6 models (the molecule’s efficiency and its toxicity for 5 animal species), to satisfy simultaneously the 6 objectives with the highest possible degree: a new molecular profile that is an herbicide having the lowest toxicity for all species.
INDUSTRY / 4iR / R&D | Mar 2023

Prediction of Concrete Compressive Strength

  • Discovery of the behavior of concrete
  • from its age, formulation and some of manufacturing characteristics
  • to predict its compressive strength and thus the level of its degradation, thanks to an explainable and accurate automated diagnosis
  • Discovery of the functional behavior of an underwater pipe
  • from its composition and physical characteristics
  • to predict its rupture thanks to an explainable and accurate automated diagnosis
  • Discovery of the highly complex hidden behavior of compressor functional dysfunction
  • from the turbine parameters
  • to make an explainable and accurate automated decay prevention
  • Discovery of effective emergency braking strategies
  • from different driving situations
  • to design an EBA that assists the driver by triggering automatically the ABS, while being able to explain each decision offline
Edit Template
DEFENSE / CYBER / SECURITY | Mar 2023

Passive Magnetic Identification of Land Mines

  • Discovery of strategies to detect a mine and identify it as anti-personnel mine or anti-tank mine
  • from 3 parameters: soil type, magnetic distortion value of the sensor, its height from ground
  • to enhance mine passive detectors (manual or autonomous) with a classifier making explainable and accurate automated decisions
  • Discovery of strategies to identify risky behaviors linked to criminal activities threatening Homeland Security
  • only from analysis of communication metadata
  • to make an explainable, automated and accurate diagnosis and to prevent attacks
DEFENSE / CYBER / SECURITY | Jan 2023

Identification of UAV Intrusion Based on Wi-Fi Analysis

  • Efficient identification of a type of invading Unmanned Aerial Vehicle (UAV) in a civilian environment
  • only from Wi-Fi traffic data records
  • to make an explainable, automated and accurate diagnosis and to prevent attacks
DEFENSE / CYBER / SECURITY | Oct 2022

Log-based Identification of Cyber Intrusions

  • Discovery of strategies of different types of cyber intrusion
  • from features of connection logs
  • to make an explainable, automated and accurate diagnosis and to prevent attacks
DEFENSE / CYBER / SECURITY | Oct 2022

Acoustic Detection of Underwater Mines

  • Discovery of strategies to detect mines
  • from sonar echoes
  • to equip vessels, submarines, and drones with a detector making explainable and accurate automated decisions
DEFENSE / CYBER / SECURITY | Sep 2022

Log-based Detection of Cyber Intrusions

  • Discovery of cyber intrusion strategies
  • from features of connection logs
  • to prevent attacks and understand the attacker's procedure
Edit Template
  • XTRACTIS robots explore a 27-dimension decision space of possible supply chain configurations to find the one that maximizes at best Operating Income given requests on local sales objectives, local production constraints, and ecological requirements. Objectives and constraints are either fuzzy (flexible) or binary (rigid).
  • Discovery of the strategy behind credit card scams
  • from a few characteristics of each transaction
  • to instantly detect the fraudulent ones and to eventually adapt protection measures

Other public Use Cases are coming soon! 

Edit Template