AI for Engineers in Simulation and Product Development
Learn about the current AI technologies, emerging trends, and how to apply AI to your simulation tasks. This training is offered as a self-paced eLearning course.
Duration
2 days
- Understand how AI technologies work in the field of simulation
- Learn to independently solve AI tasks using open-source libraries
- Gain insight into effective applications of AI
- Discover the latest trends in AI and simulation
Description
Artificial intelligence is being applied in more and more areas of our lives, including virtual product development. However, few engineers have received comprehensive training in this field, making it difficult to assess opportunities and suitable applications of AI. This software-independent course bridges this gap by linking the fundamentals of machine learning with practical exercises based on open-source libraries, thus opening the door to this megatrend. From the structure and training of neural networks to sensitivity analyses and probabilistic machine learning models, this course prepares you to understand and use modern AI tools in Ansys (such as Stochos). These tools allow you to replace simulations with trained AI models, making predictions for new designs with comparable accuracy in a fraction of the time.
Whether you are a fluid mechanic, structural mechanic, or electromagnetics engineer, this course is suitable for any engineer with basic Python skills, regardless of their physical work area.
Get a first impression and test the first eLearning module of this training course without any obligation. No costs, no notice period.
Detailed agenda for this 2-day training
Day 1
01 Introduction to Machine Learning
- Qu'est-ce que l'apprentissage automatique ?
- Quelles sont les conditions préalables à prendre en compte ?
- Obtenir un aperçu des différents types d'apprentissage.
- Examiner des exemples de domaines d'application et des avantages.
- Examiner les perspectives : utilisation de l'intelligence artificielle dans Ansys avec Stochos.
02 Overview of Simple Methods
- Fundamentals of machine learning methods illustrated with linear regression.
- Extension to nonlinear regression.
- Determining model quality and model validation.
- Exercise: Training linear regression models using example data.
03 Neural Networks I
- History, structure, training, and prediction.
- Overview of important parameters for optimal training.
- Overview of different types of neural networks.
- Exercise: Setting up and training neural networks using open-source libraries.
04 Neural Networks II
- Modeling temporal and spatial data.
- How to process image information.
- Overview of geometric neural networks.
- Exercise: Setting up and training neural networks for time and image data.
Day 2
05 Introduction to Probabilistic Machine Learning Models
- Introduction to statistical fundamentals: distribution functions and likelihood estimation.
- Introduction to Gaussian processes: the most probable function for your data.
- Overview of kernel engineering and covariance matrices.
- Exercise: Training a Gaussian process using example data.
06 Optimization & Design of Experiment
- Classification of optimization problems and overview of optimization methods.
- Advantages and disadvantages of local and global optimization methods.
- Adaptive experimental design and optimization using Bayesian optimization.
- Exercise: Optimization using adaptive methods and machine learning.
07 Sensitivity Analysis
- What is the benefit of a sensitivity analysis?
- Overview: types of correlations.
- Overview of methods to calculate sensitivities.
- Correlation coefficients for variable preselection.
- Variance-based sensitivity analysis to capture all types of correlations.
- Exercise: Identifying key influencing factors using open-source libraries.
08 Dimensionality Reduction
- Why should dimensions be reduced?
- A simple and effective method: principal component analysis explained.
- Dimensionality reduction using neural networks: autoencoders.
- Exercise: Applying various dimensionality reduction methods.
Your Trainers
Dr. Kevin Cremanns
Placement in the CADFEM Learning Pathway
Participant data
Additional information
Commentary
Get a first impression and test the first eLearning module of this training course without any obligation. No costs, no notice period.
Whether eLearning, classroom courses, live online training or customized workshops - together we identify the best option for you.
Do you have questions on the eLearning course?
If you book through your university, you will receive a 50% discount on the stated fee on training courses and eLearning courses.
For more information on the validity and how booking with the code ACADEMIC50 works, please visit our page on training for academic users.
To get a clear impression of our online learning format, we offer you a trial allowing you access to the starting module of an eLearning course of your choice. No costs, no cancellation period or anything similar. Moreover, with this free test access you can check all the technical requirements for a smooth learning process. You can easily request the free module from any eLearning course.
Each online course day comprises four eLearning modules. You should ideally allow 90 to 120 minutes of uninterrupted learning time for each module. This will allow you to acquire the knowledge provided by a module and to consolidate it through quiz questions and Ansys exercises. By dividing each module into micro learning units, you can also make good use of smaller time windows, such as on your commute.
Prerequisite for the use of the eLearning courses is the use of a personalized access to the CADFEM learning platform.
When purchasing an eLearning course, access to the learning platform is 365 days.
As a subscription user, access to the learning platform starts and ends with the start and end of the subscription.
With the purchase of a further learning product (learning subscription, training, eLearning), you will receive renewed access to all previously booked content for 365 days, starting from the start date of the new learning product.