Presentation

Overview

Recently,  there is an increased interest in data management methods.   Statistical machine learning techniques, empowered by the available pay-as-you-go distributed computing power,  are able to extract useful information from certain data.  The international press, being specialized or not, has echoed these remarkable results as a new Spring for Artificial Intelligence in a broad sense.  The data is sometimes even referred to as the “gold of the 21st century”.  In any area of business and science, one tries to construct huge datasets to be able to profit from the benefits of the Artificial intelligence revolution.

Our team is working on some specific research challenges in data management and artificial intelligence.  These include

  • Crowdsourcing: Crowdsourcing is a key enabler of the AI revolution. We address challenges that could open up new perspectives in crowdsourcing. We work on task affectation, task orchestration, skill modeling, and other issues related to crowdsourcing.
  • Human-in-the-loop methods: Crowdsourcing is a specific situation where humans are involved in data collection. We could have the involvement also in knowledge extraction or in the recommendation process. We are interested in problems involving interactions with the users.
  • Analysis and management of large graphs: Large graphs arise in various application contexts, explicitly (for example in social networks, connections between IoT devices etc.) or implicitly (structures in a language, etc. ) We work on problems on managing large graphs, specifically on storing and querying temporal networks. Also, we work on questions of analyzing these networks: we would like to understand how to describe the evolution of large networks and how to represent them to be able to answer relevant questions about the graphs.
  • Machine learning and databases: Databases are widely used to store data, however when it comes to questions of machine learning and information extraction, often the data is first extracted from database systems. We think that there could be better strategies and one should be able to realize machine learning tasks inside a database system.
  • Data science with uncertain data: We develop new data science techniques for problems where the data that we analyze is uncertain.
  • Fairness and Privacy: While extracting information from data is crucial for many applications, being able to extract information is undesirable in other cases, specifically in situations involving private or sensitive data. Also, the use of machine learning in various applications raises the question of fairness. Our work also tries to address these problems.

 

Our work has a number of applications for example, in AI-supported education, citizen science, and direct industrial uses for example, for recommendation systems, for storing IoT device connections, satellite image analysis, and others.

You are interested in working with us?

 

We are hiring, see the job offers page.

 

 

 

 

logoirisa2010

Comments are closed.