Datafari is a big data open source enterprise search solution. It can be used to index your hard drive, as well as the file shares of your organization, but also data from the web. Many crawling connectors make it multisource (Sharepoint, SMB, DBs, Web…). The difference with Elasticsearch? It is a layer above: Datafari is not only a search engine, it also provides a framework for crawling, it manages security, it provides a web interface, as well as a complete administration interface. In this initiation workshop, we will be presenting the tool and its architecture, and then we will dive directly into its installation and usage: installation of Datafari CE on your machine, first indexing and configuration, customizing of the UI. With this, you can go back home with a technology that is easy to install and providing you with a search that is multisources and multiformats.

prerequisites:

participants must come with a laptop with at least 8GB of RAM. Preferences, the laptop must be under Linux to be able to do the installation phase. If the laptops are on windows, we will skip the installation step.

The purpose of this workshop is to provide a comprehensive overview of the Machine Learning Open Studio (MLOS) platform from ActiveEon. It empowers data engineers and data scientists to quickly create structured workflows. MLOS provides a rich set of generic machine learning tasks that can be connected together to build powerful machine learning workflows for various aplications such as: fraud detection, anomaly detection, image recognition, etc. MLOS is ready-to use open source machine learning and deep learning libraries, such as scikit-learn, PyTorch, TensorFlow, Keras, CNTK, among others.

Prerequisites:  
Participants must come with a laptop with at least 8GB of RAM. 

Application developers making use of open source components need to know whether those components comprise security vulnerabilities, understand their risk in the particular application context and mitigate accordingly. This vulnerability management can be automated using an open source solution called vulnerability-assessment-tool, developed by SAP and open sourced in 2018. This solution requires a vulnerability database with detailed information about the vulnerable source code and the respective security patch, all of which is difficult to obtain due to non-reliable standard sources of advisories and vulnerability data (such the National Vulnerability Database, NVD). In our previous research work, we treated source code changes as documents in natural language processing, potentially ignoring the structured nature of source code, in order to populate the database. We will give a brief overview of the directions of our most recent work, where we incorporate the semantic properties of code into our analysis by leveraging on state-of-the art approaches to generate distributed code representations by analyzing and aggregating paths extracted from the abstract syntax tree of the code.

In this talk, we will discuss some of the current trends in AI and see how these can be applied within the travel industry. From choice modelling to recommender systems, reinforcement learning to data generation, we will showcase how these techniques can be leveraged to improve the travel experience.

Low Level 3D Embedded Graphics with Linux, OpenGL ES and Buildroot This workshop will introduce how embedded 3D GPUs are working, and how they can be programmed at low-level on a Linux systems. Live coding examples will be presented: GL ES shader programming, full embedded system build with Buildroot, live camera image processing with V4L2 and GStreamer…

Prerequisites:

People interested to do the coding examples should come with a laptop running a working Linux system (e.g. Debian 9, Ubuntu 18, Fedora 29, …) with some free disk space (about 20GB). This is a technical session, in which many programming examples will be shown.

Attendees may follow the presentation without any equipment. All the presented materials and code will be published, so it can be reproduced later on.

Attendees which would like to do live coding session, will need a laptop with Linux (tested with Ubuntu 18). 20GB of free disk space is needed for running all demos.

Prerequisites:

A laptop, with a functional docker environment and a java development environment.
We can use Kafka with other languages (javascript, python) but java is still the main language in the kafka ecosystem.

Design my Connected Object in the biggest FabLab of the Côte d’Azur: Design, prototyping, realization (3x1h)

The rise of IoT and smart infrastructure has led to the generation of massive amounts of complex data. Traditional solutions struggle to cope with this shift, leading to a decrease in performance and an increase in cost. In this talk, we will take a look at this kind of data coming from real-world smart factory sensors. Participants will learn how to create a data pipeline for ingestion and visualisation. By the end of this session, we will be able to set up a highly scalable data pipeline for complex time series data with real time query performance.

Résumé intervention

Due to the versatility of the markets addressed, operational agility is essential to industrial success. This agility has a cost in terms of complexity and heterogeneity of information and control systems. Moreover, parallel to the deployment of automation, we experience an evolution in the manufacturing trades: products are manufactured in lesser quantity and larger variations. Given, this complexity, the human being finds more and more its place as a supervisor: it guarantees the smooth functioning of the system as a whole by being able to unlock as quickly as possible any abnormal situation. Still, AI and rule-based systems prove to provide interesting « complements » to help them focus on the real problems and their possible causes. By attempting to eliminate false positive alertings, the real pain of automatic supervision, we can provide truely useful « assistants » that help reduce stress. Hurence, a recognized specialist in Big Data systems, is setting up an hybrid system, mixing rules and supervised learning, that: