You are here : Home > AI is the heroine of CEA-Leti's first web series!

Tutorials on Artificial Intelligence

Published on 15 January 2021

AI is the heroine of CEA-Leti's first web series!


If you're a fan of web series and eagerly looking forward to new releases…Let your curiosity prevail and treat yourself to CEA-Leti and CEA-List first web series broadcast on the Institute's YouTube channel! In this 2-season, 12-episode series, two CEA-Leti research engineers invite you to discover on-board artificial intelligence: edge AI.

CEA-Leti's onboard Artificial Intelligence team is preparing the future by combining innovations in architectural fields dedicated to AI applications: non-volatile memory technology, sensors or advanced algorithms such as incremental learning. By watching this web series, the layman can learn about or extend his/her knowledge of AI from its inception to the principles and applications of today's edge AI


Based in Grenoble, which was selected and labeled by an international jury and the French government as one of the four "Cities for Artificial Intelligence", CEA-Leti is a privileged partner of the MIAI Grenoble Alpes (Multidisciplinary Institute in Artificial Intelligence). This institute plays a major role in Artificial Intelligence research, teaching students and professionals, and supporting major companies in innovation.

                

Subscribe now to the CEA-Leti YouTube channel 

​Season 1 includes eight episodes and is presented by research engineer Frédéric Heitzmann, member of CEA-Leti's edge AI program.

​#1. A brief history of Artificial Intelligence

#5. Neuronal networks capacities

​#2. Machine Learning : learning from data

#6. Neural network architectures

​#3. Why dataset are so important?

#7. Learning process

#4. Models for artificial neurons

#8. More than just neurons: reinforcement learning & Generative Adversarial Networks


Season 2 is presented by CEA-List research engineer Alexandre Valentian, an AI program team member specialized in integrated circuit design. It's five episodes address four different subjects.

#1. Deep learning methodology

#3. Deep learning: learn, troubleshooting

​#2. Deep learning: learn, popular software frameworks

Each episode is in English and runs for six to eleven minutes. This simple, clear web series succeeds in sharing solid scientific content, while remaining accessible to a wide audience.

​Season 1


#1. Brief history of Artificial Intellgence


This video is the first of a series on artificial intelligence. It deals with the fundamentals of AI, about artificial neural networks, but also about the following practical problems: how to build a neural network, how to debug a neural network, how to organize your data, and so on. This video aims to explore the different existing software frameworks and the hardware platforms.

 


#2. Machine Learning - learning from data

This video is the second in a series of talks on artificial intelligence. In the previous one you learned about the history of AI and machine learning. This video will present the different methods for a machine to learn something, and there are differences between these methods.

 



#3.Why datasets are so important ?

This video is the third in a series of talks on artificial intelligence. In the previous one you learned about the machine learning and what we can learn from data . This video will present examples and discussion around problems of bias and consistency.

 



#4. Models for artificial neurons

This video is the fourth in a series of talks on artificial intelligence. In the previous one you learned about the importance of datasets, as well as examples around problems of bias and consistency. This video will present models for artificial neurons. The basic element of a neural network is … the neuron of course!

 



#5. Neuronal Networks Capacities

This video is the fifth in a series of talks on artificial intelligence. In the previous one, you learned about models for artificial neurons. This video will present the role of these neurons, its power when arranged in large networks.

 

#6. Neural Networks Architectures


This video is the sixth in a series of talks on artificial intelligence. In the previous one, you learned about the role of these neurons, its power when arranged in large networks. This video will present how it is done inside the network and the neural network architectures.


 

#7. Learning Process


This video is the seventh in a series of talks on artificial intelligence. In the previous one, you learned about neural network architectures, its power when arranged in large networks. This video will present the learning process.


 



#8. More just than neuros : Reinforcement learning & GAN

This video is the eighth in a series of talks on artificial intelligence. In the previous one, you learned about learning process. This video will present reinforcement learning & Generative Adversarial Networks.


 


​Season 2

#1. Deep Learning Methodology


​This video is the first in a series of talks on artificial intelligence. it deals with the meaning of deep learning application, the way a model is learnt and the use of learning frameworks. Let's get through deep learning methodology by definig the problem and precising data preparation and collection.

 


#2. Deep Learning: learn, popular software frameworks 


This video is the second in a series of talks on artificial intelligence. This 2nd episode presents the deep learning strategies and the 5 most popular frameworks. Explanations and recommendations for beginners, professionals and experts.
 

#3. Deep learning: learn, troubleshooting



This video is the third in a series of talks on artificial intelligence. This episode features Alexandre Valentian who takes a look at the learning phase itself and reviews troubleshooting strategies and solutions.

 


#4. Hardware plateforms for deep learning applications



T​his video is the fourth and the last in a series of talks on artificial intelligence. This episode features Alexandre Valentian who presents the various hardware platforms for deep learning applications, including CPUs, GPUs, TPUs and FPGAs