Artificial Intelligence and Machine Learning for PSM

This Master Class has been postponed and will be held in 2021. Please check back shortly for more details. 

This Master Class will give you an introduction to AI-Machine Learning and Python and its relevant libraries (Pandas, Scikit-learn, TensorFlow).
You will apply these tools through two practical examples: How to design your own fake-news detectors and how to create a face-recognition system. At the end of this Master Class, you will be better equipped to build your own Machine Learning models and contribute to projects within your own broadcaster.
 

Who it's for

  • Engineers from PSM who wish to develop in-house machine learning capacties
  • Software developers

Skills learnt

Coding: Learning python, scikit-learn and pandas to develop AI and machine learning tools
Data: How to use data to train a model and then test it
Innovation: Using tools in machine learning to develop applications that are tailor-made to the needs of PSM. Fake news and face recognition.

Course objectives

  • Understand the principles of machine learning and how to build test models
  • Acquire concrete tools to develop machine learning applications using Python and Scikit-Learn-Keras-TensorFlow
  • To apply these tools on relevant projects for PSM: Fake news detection and face recognition
  • To join a community of developers willing to create new multinational AI projects relevant for PSM

Schedule:

09:00 to 17:00 on all three days.

Content Outline:

Day 1: Introduction to AI and machine learning

  • Principles of machine learning
  • Practical case: train and validate your model
    • Training set - validation set - test set
  • How to choose a model
    • Bias and overfitting
  • How to evaluate the model
    • Performance criteria
  • Linear regression, regularization
  • Application: audience prediction, movies score prediction
  • Tools: python, scikit-learn, pandas

Day 2: Natural language processing applied to fake news detection

  • Content based fake news detection pronciples
  • Word extractions: bag of words, N-grams
  • Word importance meaurement: TF/IDF
  • Word embedding, BERT
  • Design classifiers to detect fake news
  • Tools: python, scikit-learn, NLTK, Word2Vec, BERT

Day 3: Deep learning for face recognition

  • Deep learning principles
  • Embedding for face recognition
  • Train and test your model on real datasets
  • Tools: python, tensorflow, scikit-learn

Equipment needed:

Personal computer

 

 

Formateur

Skill

Coding Data Innovation

Détails

22 - 24 sept. 2020

Physical

Lieu

European Broadcasting UnionL'Ancienne-Route 17A Le Grand-Saconnex GE 1218 Switzerland

Date limite d'inscription

18 sept. 2020

Détails concernant l'inscription

Fee:
EUR 1450
The third participant from the same organization will be free.

Contactez Nous

EBU Academy does not make hotel reservations. However, we have negotiated competitive rates with some of the hotels near the venue and the city centre.

Please contact the hotel of your choice directly.

Hotels offering special rates for the EBU/Eurovision (pdf)

Langues de travail

English

Partenaires et Sponsors

This Master Class has been developed in collaboration with EBU Technology & Innovation

Contacts

Frederic Frantz
Responsable Développement - Formation des Membres
+41 22 717 21 48
frantz@ebu.ch
LinkedIn