On data science environments
This tutorial provides hands-on coding experience in training, tuning and designing deep learning (DL) data centric applications. It will particularly compare large scale processing environments with respect to underlying data management issues. The tutorial will provide a concise and intuitive overview of the most important aspects of training and application of DL models to label data sets (e.g., images).
The experience of training and deploying DL applications considering data management underlying issues will be done through hands on experiences on top of the data science Databricks solution deployed on a public cloud accessible to students. The tutorial will finally compare different cloud-based data science environments, like AWS, Google Cloud AI, MS Azure AI, and IBM Watson.
Genoveva Vargas Solar (http://www.vargas-solar.com) is a senior researcher of the French Council of Scientific Research (CNRS) and she is member of the HADAS group of the Laboratory of Informatics of Grenoble, France. She contributes to the construction of service based database management systems. The objective is to design data management services guided by Service Level Agreements (SLA) to provide methodologies, algorithms and tools for integrating, deploying and executing a service oriented data management functions. Quality of service criteria and behaviour properties included in SLA are adapted to applications requirements (e.g., transactional execution, reliability, fault tolerance, evolution and dynamic adaptability). She conducts fundamental and applied research activities for addressing these challenges on different architectures ARM, raspberry, cluster, cloud, and HPC.
Jorge García Flores (https://lipn.univ-paris13.fr/~garciaflores/) is a research engineer at the French Council of Scientific Research (CNRS) at the LIPN-Lab at Université Paris 13. His research field focuses on Deep Learning engineering, information extraction, natural language processing and digital humanities.