Subsequence similarity search is an important problem, with many applications in several diverse domains. In this project, we focus on seismic signals, and address the problem of fast and scalable search for similar seismic subsequences.

PLATON (Platform-aware LArge-scale Time-Series prOcessiNg) brings together a highly-experienced researcher in the field of the theory of concurrent and distributed computing with a hosting group which has world-leading expertise on data series management, indexing, and analysis, to harness the difficulties of large-scale data series processing by realizing the data series processing performance and scalability goals. Specifically, PLATON aspires to build, for the first time, the necessary methods, algorithms and tools for highly-efficient, scalable, and fault-tolerant processing of huge collections of data series. Scalability will be accomplished by exploiting the full computational capacity (multiple nodes, multiple cores, accelerators) of modern computing platforms. To ensure fault-tolerance, PLATON is designed to run in a way agnostic to the number and identities of the computing elements provided by the platform. PLATON demonstrates its value proposition using two real datasets from different domains. The proposed research project has the potential of great economic and social impact in Europe as multiple scientific and industrial fields are currently in need of the right tools, in order to handle their massive collections of data series, a fact acknowledged by the [EU strategy on Big Data](https://ec.europa.eu/digital-single-market/en/towards-thriving-data-driven-economy).

The ANR funded "QualityOnt: High Quality Knowledge Graphs from recent English, French and German Emergent Trends with the example of COVID-19", is a joint research project with the University of Lübeck and Université Paul Sabatier, that uses knowledge graphs as the basis for the automatic analysis of the COVID-19 pandemic. The goal of the project is to combine data quality assessment criteria coupled with natural language processing techniques in order to guide the process of knowledge graph enrichment.

Sequential data are everywhere, from DNA sequences to astronomical light curves, and from aircraft engine monitoring data to the prices of stock options. Recent advances in various fields such as those of data storage, networking and sensing technologies, have allowed organizations to gather overwhelming amounts of sequential data at unprecedented speeds. This wealth of information enables analysts to identify patterns, find abnormalities, and extract knowledge.
Nevertheless, current relational storage layers cannot handle the access patterns that analysts are interested in, without scanning large amounts of unnecessary data or without large processing overhead. Thus, making complex analytics inefficient. In order to exploit this new opportunity, we plan to develop specialized data series storage and retrieval systems, which will allow analysts – across different fields – to efficiently manipulate the sequences of interest.
The proposed research project, named NESTOR (Next gEneration Sequence sTORage), has the potential of great economic and social impact in Europe as multiple scientific and industrial fields are currently in need of the right tools, in order to handle their massive collections of data series.

This project is supported by a European Union Marie Sklodowska-Curie Individual Fellowship (IF).

The aim of the project is to realize a novel service for Event Managers to sense their City Scale Events by fusing a variety of volatile and massive data available in digitized cities such as: mobile telecom data (GSM, GPRS, UMTS, and LTE); social media streams from multiple location-aware social networks (twitter, instagram, and foursquare); and pollution, traffic, public transport status data. The expected output is an interactive dashboard that allows Event Organizers to assess several conditions related to events: how people responded to their City Scale Events (number of participants, demographics of participants, etc.); how participants perceived the event (sentiment expressed with regards to the event topic); and the impact the event had on the city (anomalies in traffic, pollution, public transport, etc.).

This project is supported by the European Institute of Innovation and Technology (EIT) ICT Labs.

The falsification of medicinal products and the online trade in medicines are two closely intertwined phenomena reaching alarming proportions and having an increasingly detrimental impact on public health. Internet usage enables criminals to sell potentially dangerous products on a large scale, directly to buyers, whilst circumventing the legal supply chain. This has largely been facilitated by the emergence of so-called online pharmacies since the 1990s. The project will improve and systematise knowledge on the demand and supply mechanisms by using traditional and innovative research methods, such as virtual ethnography, Web surveys, direct observation of clients’ behaviour on ‘honey-pot’ websites, and data mining; create specialised figures through advanced training courses dedicated to LEAs; develop an ICT tool prototype for the automatic detection of illegal/rogue online pharmacies (for use by LEAs); improve awareness-raising campaigns; and develop a tool to determine the risk that a given medicine may be falsified and traded online.

This project is supported by the European Commission.

Manufacturing is the driving force of Europe's economy, providing over €6,553 billion in GDP. The KAP project will deliver energy management standards and a technology framework for next generation, sustainable manufacturing. KAP stands for Knowledge of past performance, combined with Awareness of the present state, which together can support Prediction of future outcomes. This philosophy forms the basis of a framework that will enable every existing resource to be used as efficiently as possible through the effective co-ordination of man, machine, material and method. Measurements will be gathered through a factory-wide network of sensors. Complex Event Processing (CEP) and data stream analysis will compute on-the-fly production performance indicators (PPIs) for real-time monitoring. Data mining in combination with OLAP will support problem diagnosis and resolution. Computational learning techniques will create a self-improving system for operational control.

This project is supported by the European Commission FP7 Framework Programme.

There is an increasingly pressing need, by several applications in diverse domains, for developing techniques able to index and mine very large collections of sequences, or time series. In our work, we will design techniques for indexing and mining truly massive collections of time series, focusing on the task of predictive maintenance for electric power plants. Existing techniques to identify statistical precursors of imminent critical transitions in the system under consideration are, due to the computational complexity of their traditional implementations, limited to comparably small sets of time series. One of the key foci of this proposal will therefore lie on the development of efficient, scalable algorithms to identify such precursors.

This project is supported by the Jacques Hadamard Mathematical Foundation (FMJH).

This project is investigating data quality, where the acquisition mode is diverse and involves uncertainty. The study will develop techniques for pre-processing, repairing and extraction of knowledge, in order to better analyse and highlight them for decision making. The data used are relevant to international migration. The aim is to enhance the social and political concerns for the mobility of minorities.

This project is supported by CNRS Mastodons.

In ths project, we study the quality of big data through an innovative approach to virtually clean and transform data from life sciences, guided by queries. We are also interested to handle uncertainty of data, as well as confidentiality. We use clinical and microscope photonics data.

This project is supported by CNRS Mastodons.

Time series, also known as light curves, can be used to distinguish objects in large data sets and thus to derive clean samples. This can be applied, for example, in the case of the large sky surveys of Euclid and LSST, in order to eliminate super massive black holes (AGN) from the sample of galaxies. By applying time series similarity matching, we can develop novel and effective ways for distinguishing objects of different variability behavior, leading to more accurate and higher quality overall results.

This project is supported by CNRS Mastodons.

Predon is a CNRS Mastodons project, dealing with long-term scientific data preservation.

In this project, we study workflow management for scientific data preservation, with applications on particle physics. A large class of scientific experiments are in fact unique, because of their large scale, with very small chances to be repeated, or superseded by new experiments in the same domain: for instance high energy physics and astrophysics experiments involve multi-annual and even multi-decades developments, unlikely to be repeated. Other scientific experiments are unique by nature: earth science, medical sciences etc., since the collected data is “time-stamped” and thereby non-reproducible by new experiments or observations. It is therefore of outmost importance to pursue a coherent and vigorous approach to preserve the scientific data at long term. The preservation remains nevertheless a challenge due to the complexity of the data structure, the fragility of the custom-made software environments, as well as the lack of rigorous approaches in workflows and algorithms. One of the main missions of this project is to enforce the efforts to preserve the scientific data in France.

More information on the PREDON project.

Life Imaging is an interdisciplinary project of the Paris Sorbonne Cité consortium of Universities, related to medical images management and processing.

The aim is to create an intelligent atlas of medical images. We are focusing on linking multimodal and heterogeneous medical images, using new browsing approaches, i.e., linked data and crowdsourcing.

More information on the Life Imaging project.

Square Predict is a French National project under the call projet Investissement d'Avenir Cloud Computing/Big Data (Fonds national pour la Société Numérique (FSN)), on big data analytics for insurance companies.

The project will allow to collect, integrate, and mine big data for insurance companies, while preserving the privacy of the clients. In particular, we study data quality issues in big data, including the topics of uncertainty, reasoning with inconsistency, semantics in big data, and privacy.

This course is for students in Licence 3 Informatique, as well as in Master 1 Informatique.

It focuces on the interoperability of information systems and service-oriented architecture (SOA), and covers the topics of IS interoperability, business process design, REST architecture, SOA, SOAP web services, and RESFul Web services.

This course is for students in DUT1/2 Informatique.

It focuses on relational database design and coding using the Ms Access and Oracle database systems, and covers the topics of data conceptual modeling, relational model, sql, query language, and server-side programming (pl/sql, triggers).

This course is for students in Master 1

Informatique. It covers topics relevant to data analytics for large scale data. We start by discussing methods and techniques for data pre-processing (data integration, data cleaning, data summarization), and we then present data analysis techniques (frequent itemsets, clustering, classification) and how they have evolved in response to the requirements of large scale and streaming data analytics. Students participate in group projects, and get hands on experience on the subjects presented in the class.

rerequisites: good knowledge of databases, good programming skills, basic knowledge of probability and statistics.

This course is for students in Licence 3 Informatique.

It is the second course on databases in the undergraduate program, and covers the topics of data storage (disk, memory, buffer pool), file organization (heap files, page and record formats), data indexing (B+-Trees, Hash Index), implementation of database operators (external sorting, selection, projection, join), and query optimization (query plans, cost estimation, histograms).

Students get hands on experience on the subjects presented in the class.

Prerequisites: good knowledge of databases, good programming skills.

This course is for students in Master 1 Miage Initiale and Master 1 Miage Apprentissage.

It covers topics related to database management systems including storage, indexing techniques, query processing and optimization, transaction management, and distributed databases. It also introduces emerging database-related techniques.

Prerequisites: good knowledge of databases.

This course is for students in Master 2 Miage Initiale and Master 2 Miage Apprentissage.

It covers the topics of the knowledge discovery process, data warehouse architectures and modeling (ETL, data marts, MOLAP, ROLAP, star schema), and key data warehouse concepts (OLTP, OLAP, datacube, bitmap and join indexes, warehouse granularity, incremental maintenance, distributive and algebraic functions).

Students participate in group projects, and get hands on experience on the subjects presented in the class.

Prerequisites: good knowledge of databases, good programming skills, basic knowledge of probability and statistics.

This course is for students in Master 1 Informatique.

It covers topics relevant to data management and data analytics for Big Data. We start by presenting some well-established data management and processing techniques (data organization, data indexing, cost-based optimization), and we then discuss how these techniques have evolved, or changed, in response to modern hardware and architectures, as well as to the requirements of Big Data.

Students participate in group projects, and get hands on experience on the subjects presented in the class.

Prerequisites: good knowledge of databases, good programming skills, basic knowledge of probability and statistics.