Artificial Intelligence in Mechanical Engineering

Artificial Intelligence in Mechanical Engineering#

by Marc BUFFAT, dpt Mécanique, Université Lyon 1[1]

_images/logo.png

Initial reflexion#

Moravec’s Paradox (Roboticist) :

_images/IAquestion1.png

Fig. 1 (C) France Culture 2019: l’intelligence artificielle en question#

This paradox deals with the often mistaken fact that it is the logical problems in the field of Artificial Intelligence that are the hardest to solve, and that simple day to day activities like facial recognition, and hand-eye coordination that are the easiest and can be easily implemented.

What it means is that day to day activities like speech perception, facial recognition, and motor activities are actually harder to implement in Artificial Machines than making a machine play chess or carry out any other activity that requires logic and computation.

Introduction and E-book content#

The objective of this course is to be able to use a machine learning approach applied to the treatment of mechanical problems. The adopted perspective is to present the machine learning approach from a numerical methods point of view (using applied numerical methods) rather than from a statistical point of view, which is mainly used by data scientists.

This E-book is divided intro 2 parts:

Part 1 Course Content#

_images/IA_mechanic1.png

Introduction to Machine Learning#

  • Introduction to AI

  • Introduction to Machine Learning

Mathematical Basis of AI algorithms#

  • Minimization methods for Machine Learning

    • Minimization for machine leraning

    • Lecture notes for students

Overview of Machine Learning Algorithms#

  • AI Algorithm for supervised regression

    • Main AI Algorithms

    • Lecture notes for students

Use of AI Libraries#

  • Methodology and use of AI libraries

Time series data processing for machine learning#

  • Recurrent Neural Networks: RNN and LSTM

    • Time series

    • Lecture notes for student

AI/ML Computing#

  • Software is the key of AI, but knowledge of the underlying hardware is extremely important to run the software efficiently

    • multi-threding

    • use of GPU

Part 2: Practical works#

  • Practical work: introduction

  • Practical work: prediction of mechanical properties

  • Practical work: machine learning for shock propagation

  • Prediction of the movment of the double pendumum

  • Pratical work: meteorological prediction

Table of Content#

Part I Course content

Part II Practical works

Annexe

Contents