Practical A.I. & Deep Learning for Professionals

So, you’ve read about ChatGPT and other generative A.I. recently. But did you know underneath is a rich history of Deep Learning algorithms that are still useful today? This course is a hands-on practical experience that we specifically curated for industry professionals aiming to leverage the power of Artificial Intelligence (AI) and Deep Learning in their respective fields.

We aim to demystify complex AI concepts, algorithms, and tools, and translate them into actionable insights and practical applications.

IS IT THE COURSE FOR ME?

Do you want to learn to work with AI/Deep Learning in a practical approach that fits real-world use cases? Do you want to learn the ‘full stack’ approach, where you learn not just how to train and optimize an AI model but also deploy it to a web environment and put an API on top of it? Are you familiar with Python on at least a medior level?

Then this is the one for you!

WHAT TO EXPECT FROM THE TRAINING

Expect a highly interactive training with the goal of giving you practical knowledge. Our training day follows a practical template where we give compact (30-60 min) info sessions followed by 60-90 minutes of hands-on lab work. This way you can immediately put what you learned into practice and create lasting knowledge, rather than passively listen and forget most things again two weeks later.

Deep Learning is a complex and broad field, and as such we will take you through the various applications; you will work with image data, text data, audio data, as well as numerical data. You will work with some of the leading tooling like PyTorch, and TensorFlow/Keras.

While we do share the mathematics behind various algorithms, we will not dwell too long on them. The most important part is that we give you an intuition about what’s happening under the hood, as this will greatly increase your efficacy in implementing successful deep learning algorithms.

REQUIRED PARTICIPANT BACKGROUND

The course is given in Python, the leading data science language, and offers frequent and in-depth lab-work. In the lab-work you apply the theory you just learned into practice right away. Our exercises come in two flavors: ‘regular’ and ‘advanced’.

The ‘regular’ exercises offer you a solution that has been worked out for a large part, and asks you to ‘fill in’ various sections. For this you need to have a good understanding of Python at a medior level. The ‘advanced’ exercises offer you a barebones skeleton framework and offers you more flexibility in how you approach the final solution. This requires advanced Python knowledge.

Some additional information about the exercises: if you’re interested in getting some experience with Deep Learning but are not expecting to use it often, the ‘regular’ exercises offer a great way of dipping your toes into the water and playing with it. If you plan to use AI/Deep Learning often and/or in a professional setting we highly encourage you to work on the ‘advanced’ exercises, as these offer you more in-depth experience, and resemble real-world use cases more than the regular ones do.

BROAD OVERVIEW OF THE TOPICS COVERED

• Fundamentals, working with Python and data in an AI setting
• Getting your feet wet: supervised machine learning
• A.I. foundations and neural nets
• Convolutions as a big leap
• Advanced techniques: regularization, hyperparameter tuning, transfer learning, data augmentation, grid search, among others.
• First aid “help my model is not learning/performing”
• Memory in networks: RNN, LSTM
• Generative Adversarial Networks
• Large Language Models

You will work with:

• Image data
• Audio data
• Video data
• Text data
• Numerical data

ABOUT THE TRAINER

Dr. Paul van Gent

The course will be given by Dr. Paul van Gent. He started his career developing sensor hardware and firmware for use in road safety research at SWOV in the Hague, along with analysis tooling for the data collected. After several years he started a PhD at Delft University of Technology, during which he designed various algorithms for analyzing physiological data, including the successful package HeartPy that has since seen >100K downloads. During his PhD and following Postdoctoral work, he developed and taught a hands-on deep learning workshop and worked with Deep Learning in his research often.

Currently, he is actively involved with KPN, where he is working as a backend engineer and data scientist. Here, aside from working on new product features, he is setting up various data solutions to offer automatically generated business intelligence reports and insights to be used by critical stakeholders in the company.

Aside from his professional life, Paul is a father of one, plays the piano and composes music, plays Dungeons&Dragons as a dungeon master with friends and colleagues, and in general likes to stay active.

Extra information

This training can be given in Dutch or English, based on the language of the participants.