World Artificial Intelligence Cannes Festival 2023

We are exhibiting at BOOTH A13.

See the program

Learn more about the demos

Program

BONUS: Rage against the machine learning - Concert provided by 3IA Côte d'azur on Friday 10th from 6 to 7 pm!

3IA Techpool Demos

DeepWILD
Detection and monitoring of animal species in the Mercantour

Summary: Researchers, students and engineers have contributed to develop an AI algorithm capable of locating, identifying and quantifying the species present on footages.
The Mercantour national Park wishes to monitor and identify common and rare species present on its territory from videos extracted from camera traps located on the Roya Valley.
An algorithm composed of convolutional neural networks (Faster-RCNN and InceptionResNetv2) has been implemented thus gaining in efficiency to count and estimate the fauna present.

Chair of Charles Bouveyron (Université Côte d'Azur - Inria)

DispuTOOL
A tool for Mining and Exploring arguments in US Presidential Election Debates From 1960 to 2016

Summary: Political debates are the means used by political candidates to put forward and justify their positions in front of the electors with respect to the issues at stake. Argument mining is a novel research area in Artificial Intelligence, aiming at analyzing discourse on the pragmatics level and applying a certain argumentation theory to model and automatically analyze textual data. DISPUTool is a tool designed to ease the work of historians and social science scholars in analyzing the argumentative content of political speeches. More precisely, DISPUTool allows to explore and automatically identify argumentative components over the 39 political debates from the last 50 years of US presidential campaigns (1960-2016).

Chair of Serena Villata (CNRS)

Fed-BioMed
Federated Learning Framework

Summary: Standard machine learning approaches require to have a centralized dataset in order to train a model. In certain scenarios like in the biomedical field, this is not straightforward due to several reasons like: privacy concerns, ethical committee approval or transferring of data to a centralized location. This slows down research in healthcare and limits the generalization of certain models.
Fed-BioMed is an open source project focused on empowering biomedical research using non-centralized approaches for statistical analysis.
Federated Learning (FL) is a machine learning procedure whose goal is to train a model without having data centralized. The goal of FL is to train higher quality models by having access to more data than centralized approaches, as well as to keep data securely decentralized. The main challenges are associated to: communication efficiency, data heterogeneity and security.

Chair of Marco Lorenzi (Inria)

ReID
Re-identification from different angles

Summary: Demonstration of supervised and unsupervised object recognition algorithms from camera images with different viewing angles.
Person re-identification (ReID) is an illustration of a fundamental problem in Computer Vision, how to represent an object (i.e. its appearance) independently from its perception (i.e. its pose, view-point, illumination and the given sensor). As a core component of intelligent video surveillance system, person ReID has attracted increasing attention in computer vision research community. Given a query person image, a person ReID system seeks the most similar images in the gallery by sorting representation similarity. Although rapid improvements has been witnessed in recent years, person ReID still suffers from real-world environmental factors, such as illumination, pose, view-point and background variance, which degrade the quality of image representations. How to build robust representations that are invariant to the environmental factors with limited data remains a key problem for person ReID. In addition, annotating a large-scale person ReID dataset is a cumbersome task, which strongly limits the scalability of supervised ReID methods.
In our work, we first study how to build robust representations for supervised person ReID. Then, we remove the human supervision and design unsupervised algorithms to enhance the flexibility of person ReID.

Chair of François Bremond (Inria)

Indago
Communication Network Analysis

Summary: Understanding communities and topics of discussion in communication networks through AI
Indago is a software for the analysis of communication networks with textual edges. It helps you to understand the different communities and topics which exists inside (very) large networks. It is built upon STBM (Stochastic Topic Block Model), a non-supervised AI model that simultaneously analyzes network connections and the content exchanged inside these connections.
With Indago, you will be able to get the most out of social media, email netwotrks, co-authorship networks, or even work with your own data.

Chair of Charles Bouveyron (Université Côte d'Azur - Inria)

EURECOM Demos

Emulating high-dimensional simulators using statistical machine learning

Summary: We present our recent approach to emulating high-dimensional simulators using statistical machine learning. We apply it to the emulation of a tsunami simulator, motivated by tsunami early warning.

Demo by Motonobu Kanagawa (3iA Chair)

Exploring the Vulnerabilities and Defenses of Principal Component Analysis on Medical ECG Data: A Membership Inference Attack and Differential Privacy Demo

Summary: This demonstration will show the performance of a membership inference attack against principal component analysis (PCA) applied to Medical ECG data. The audience will assume the role of an adversary who has access to the principal components and aims to infer whether a given data sample was used to compute these components. We will demonstrate the performance of the attack when the number of samples used to compute the principal components changes. As a defense strategy, The audience will also see the experimental results on the performance and utility of differentially-private PCA solutions when applied to Medical ECG data.

Demo by Oualid Zari (Chair of Melek Önen)

Autonomous flying robots for sensing and connectivity

Summary: The future is the era of connected robotics thanks to super-fast wireless communication. With the wave of connected intelligence, we are passionate about connecting the sky to the ground with cutting-edge communication technology (5G) and the fast-growing robotics industry. 
At EURECOM, where OpenAirInterface (OAI) the pioneer of the open source 5G/LTE implementation was born, we try to make one step closer to the future of connected intelligence. Our goal is to research and prototype autonomous flying robots powered by OAI-5G radio connectivity. We consider various scenarios ranging from AI-based flying base stations/relays to radio frequency (RF) sensing and localization applications. Our research is at the intersection of machine learning, robotics, and wireless communications.

Demo by Omid Estafilian (Chair of David Gesbert)

AI for health: Using AI to identify stress from wearable devices data

Summary: Wearable devices represent a promising solution to achieve real-time monitoring of a person’s health status beyond the clinic. Given the large amount of data that these devices can collect from a single user within a short period of time, artificial intelligence (AI) tools are necessary to analyse the data and draw conclusions. In this demo, we will showcase how we combine AI with wearables collected data for the prediction of stress.

Demo by Hava Chaptoukaev / Bianca Dalpaos (Chair of Maria A. Zuluaga)

DOWNLOAD YOUR DISCOVERY PASS TO VISIT US