16 novembre 2021 – SophI.A Master Classes by Telecom Valley

Présentation de la stratégie DATA de la collectivité ; comment dans un contexte, une collectivité comme le conseil départemental des alpes maritimes s’organise pour placer la data au coeur de son action ; Etat de la démarche, difficultés et opportunités ; Retours sur des expérimentations menées d’utilisation de l’intelligence artificielle

In this master class, we will give an overview of the different anomaly detection techniques for time series. We will focus more particularly on unsupervised, mono-varied anomaly detection techniques applied to satellite telemetry. Finally, several of these algorithms will be implemented in python (scikit-learn, tensorflow, etc.) and their performances will be compared.

Fare discrimination based on customer segmentations and deceptive choice architectures have always existed, offline or online. The online model, with the depth of its data collection, its fine-grained personalization and its pervasive AI-boosted algorithms seem to have brought the concept to another level. 

The aim of the masterclass is first to highlight the mechanisms used by artificial intelligence algorithms to discriminate and exploit customer behavior in order to reach the economic goals of digital platforms. Using concrete examples inspired by recent scandals and court decisions, we will show that detecting biases, disloyalties or fraud advocate interdisciplinary approaches, both legal and technical. A challenging task for regulation authorities. 

During this Master Class, you will have the opportunity to build your own database in a fun and interactive way for posture recognition.
You will then develop and train a neural network capable of classifying poses in order to test it on the camera provided.

For these tasks, you will complete and customize a script provided by the VIDETICS team.