Introduction to Deep Learning for Artificial Intelligence

84,00 €

In an era where artificial intelligence (AI) is at the forefront of technological and economic advancement, understanding its intricacies has become crucial for professionals across various sectors. This intermediate-level course aims to equip participants with a comprehensive understanding of deep learning, a pivotal branch of AI responsible for significant breakthroughs in the digital world. Targeted at any individuals with a basic understanding of Python programming and mathematics, the training seeks to demystify the scientific and technological foundations of deep learning, including linear algebra, calculus, and software engineering, alongside practical skills in using deep learning frameworks like PyTorch and Keras.

 

The course covers essential concepts from the basics of neural networks, activation functions, and data handling, to the application of various architectures like ANNs, CNNs and Transformers. It emphasizes hands-on learning through exercises, projects, and interactive discussions, ensuring participants can design, implement, and refine neural networks effectively. Moreover, it addresses the social and ethical dimensions of AI, preparing attendees to make responsible decisions in AI deployment or use. This blend of theoretical knowledge and practical application, set in an on-site format conducive to immersive learning, makes the course an invaluable opportunity for those looking to deepen their expertise in AI or integrate AI solutions into their work, fostering a future-ready skill set in the rapidly evolving landscape of artificial intelligence.

 

This course will cover:

  • Scientific and technological foundations
  • Basic concepts of deep learning
  • Main architectures of deep learning
  • Mini-Project – Applying the learned concepts
  • Ethics and society – Organizing to live with AI

9 in stock

SKU: 14601 Category:

Start date

June 13, 2024

End date

June 27, 2024

Language(s) of the training

English

Languages spoken by the coach(es)

English, French

Instructor(s)

Prof. Dr. GUELFI Nicolas

Contents

Discovering the Scientific and Technological foundations

  • Science
    – Linear algebra
    – Calculus
    – Probability and statistics
    – Computational models
    – Optimization & information theories
  • Algorithmic complexity
  • Software engineering
  • Technology
    – Processing units: GPU, TPU
    – Deep learning frameworks and libraries: Pythorch, Keras
    – Cloud computing platforms: GCP
    – Development tools and environment: Jupyter notebooks
  • Practical exercises with predefined objectives are followed by a feedback session on the experiences carried out, allowing for discussion and reflection on what has been accomplished.

 

Understanding the basic concepts of deep learning

  • Basic deep learning concepts
    – Neural network, parameters, hyper-parameters, activation function, batch, bias, input, output and hidden layers
    – Classification, regression
    – Datasets (training, validation, and test)
    – Efficiency of a neural network (cost function, accuracy, overfitting or underfitting)
    – Training (forward/backward propagation, gradient descent, dropout, epoch, learning rate)
  • Presentation of the example of handwriting recognition (i.e., MNIST) used in the module as a running example.
    – Practice of the basic concepts using progressive and applied exercices

 

Understanding the main architectures of deep learning

  • Basic
    – Artificial Neural Network (ANN)
    – Long Short-Term Memory network (LSTM)
  • Natural Language
    – Self-attention network (Transformers)
  • Vision:
    – Convolutional Neural Network (CNN)
    – Generative Adversarial Network (GAN)
    – Vision-Transformers (ViT)
  • Practice of the Main deep learning architectures using progressive and applied exercises on MNIST

 

MINI-PROJECT – PART 1 – Apply the learned notions of deep learning in the context of a mini-project

  • Introduction
    – Project objectives and requirements
    – Input libraries
  • Data set preparation
    – Image synthesis
  • Neural Network Architecture

 

MINI-PROJECT – PART 2 – Apply the learned notions of deep learning in the context of a mini-project

  • Round Trip engineering
    – Training
    – Results analysis
    – Architecture improvements
    – Data set improvements

 

Understanding how societies can organize themselves to live with AI

  • Ethics & AI
    – Understanding the notion of ethics in the context of Artificial Intelligence
    – Synthetic presentation of the main approaches to provide an ethical framework for the development of AI
    – Proposals from main organizations (UNESCO, OECD, Parliament/Council/Commission of Europe)
    – Presentation of the European regulation on AI (i.e., AI Act)

Objective

Practice based learning of concepts, methods and solutions.

Learning Outcomes

After completion of this training, you will be able to:

  • Design, implement, and refine neural networks, including setting and adjusting parameters like weights, biases, learning rates, and understanding the mechanics of forward and backward propagation, dropout, and data augmentation techniques.
  • Apply various deep learning architectures such as ANNs, LSTMs, Transformers, CNNs, GANs, and Vision Transformers to appropriate domains, such as natural language processing, image recognition, and generative tasks.
  • Demonstrate expertise in managing datasets, including splitting into training, validation, and test sets, and competently training models while optimizing performance and preventing over-fitting.
  • Understand the social and ethical implications of AI technologies, ensuring responsible and ethical decision-making in the development and deployment of AI solutions
  • Use Jupyter Notebooks as a versatile tool for developing, documenting, and testing deep learning models in an interactive environment.

Schedule

This training has a total duration of 21 hours and takes place over 3 days:

  • 13-06-2024: 09:00 – 17:00
  • 20-06-2024: 09:00 – 17:00
  • 27-06-2024: 09:00 – 17:00

Format and Location

This course takes place ON-SITE
Terres Rouges building
14, porte de France
L-4360 Esch/Alzette

Level

Intermediate

Prerequisites

Preliminary notions in Python equivalent to the content of the following trainings at DLH:

  • Python basics (modules P1 to P3)
  • Python – Basics Camp

Initial understanding of basic mathematics notions (algebra, calculus, statistics, probability).

Additional Info

Certification

This training does not have any assessment or exams; a certificate of participation will be issued to participants.

Esco Skills

gather data, process data, inspect data, ethics, forecast future ICT network needs, emergent technologies

Esco Occupations

ICT teacher secondary school, secondary school teacher, software developer, data scientist, data quality specialist, data entry clerk