School / Prep
ENSPIMA
Internal code
AP9MFIAR
Description
Objectives
The main objective of this module is to provide an introduction to Artificial Intelligence through one of its major components, Machine Learning (ML), and more specifically Learning Neural Networks. The main classes of problems addressed by ML are discussed, with their principles, tools and limits of use, in the field of engineering. The programming, training and operation of neural networks are illustrated from a practical point of view by programming in the Python language with the tensorflow and keras modules.
Skills acquired at master's level
Have a global systemic approach
Teaching hours
- CMLectures15h
Mandatory prerequisites
Knowledge of the Python language for data processing and curve plotting using the numpy and matplotlib modules.
Syllabus
Contents
The module comprises two 1h20 lecture sessions and 4 3h practical sessions.
The practical part of the module is structured around two sequences, preferably carried out on students' laptops:
Discovery of Machine Learning: programming, training and evaluation of a neural network (dense then convolutional) dedicated to image recognition (handwritten digits).
Application to the context of aeronautical maintenance: design, training and evaluation of a specific neural network dedicated to the detection/classification of defects in data from an engine test bench.
Target learning outcomes:
Know the main classes of problems addressed by Machine Learning.
Explain how an artificial neuron works and the overall architecture of a neural network.
Build a dense or convolutional neural network using the tensorflow and keras Python modules.
Download, prepare and use a labeled data set to train a neural network.
Train a neural network and exploit precision and loss curves to limit overfitting.
Know how to exploit a trained network (by yourself or others) using a new data set.
Teaching method
Practical work is carried out in Python language, preferably on students' laptops. Teaching progress is based on a series of notebooks (IPython notebooks) with "holes" of increasing difficulty, to help students acquire the skills they are aiming for.
Further information
Maintenance of the future
Assessment of knowledge
Initial assessment / Main session - Tests
Type of assessment | Type of test | Duration (in minutes) | Number of tests | Test coefficient | Eliminatory mark in the test | Remarks |
---|---|---|---|---|---|---|
Integral Continuous Control | Continuous control | 1 |
Second chance / Catch-up session - Tests
Type of assessment | Type of test | Duration (in minutes) | Number of tests | Test coefficient | Eliminatory mark in the test | Remarks |
---|---|---|---|---|---|---|
Final test | Oral | 1 |