Introduction: Time series analysis across disciplines

• Volker Beckmann (CNRS / IN2P3)

Room: Room Conseil (1) The study of time series is common to most scientific disciplines. Throughout the domains, similar questions are asked to the data, such as what are the trends (in time or along any other parameter) detectable in the data, is there recurrent behaviour, can correlations between parameters be established, how significant are they and are they causal? In addition, the evolution of computing power enables researchers to combine large data sets of different origin and from different fields in order to derive new insights. With these commonalities, why is interdisciplinary work in the domain of time series analysis still rather scarce? I will address some of the issues and point out developments that help us building up collaborations to benefit from each others expertise in this research area.

Data series management: first results

• Themis Palpanas (LIPADE - Paris Descartes University)

Room: Room Conseil (1)

Astronomy - photon: Fast timing variability in X-ray binaries: methods & diagnosis

• Jérome Rodriguez (CEA)

Room: Room Conseil (1) In this talk I will first introduce the astrophysical systems at the core of my research, X-ray binaries (a.k.a. microquasars), and will in particular focus on their temporal properties. I will then show how we can try to understand the mechanisms at work in these sources through a timing approach, i.e. the study of their X-ray light curve  and associated evolution on long (day) timescale. Rapid (sub-second) variability is studied through Fourier analysis which has shown a vast richness of behavior in these sources, allowing to go deeper into several domains of (astro)-physics.

Astronomy - photon: X-ray variability in accreting supermassive black holes and relations to X-ray binaries

• Gabriele Ponti

Room: Room Conseil (1) I will discuss how the variable X-ray emission from accreting supermassive black holes (or active galactic nuclei, AGN) can be used to extract information about the central compact object. I will show the evolution of the variability properties with black hole mass and accretion rate, within the sample of AGN and how this is related to variability in X-ray binaries. I will finally introduce some new fourier-time-delay techniques that carry the potential to echo-map the geometry of the emitting region.

Computer Science and Engineering

• Mueen Abdullah

Room: Room Conseil (1)

Astronomy - gravitationnal waves

• Vivien Raymond (AEI-Potsdam)

Room: Room Conseil (1)

Visualization and human-computer interaction methods: Challenges in Interactive Visual Exploration of Large Data Series

• Anna Gogolou

Room: Room Turing (2)

Neuroscience: focus on electrophysiology of epilepsy

• Katia Lehongre (Brain and Spine Institute (ICM))

Room: Room Turing (2) There are many techniques used in neuroscience to explore the brain activity, going from electrophysiology to functional magnetic resonance imaging. All the techniques used, despite various spatial and temporal resolutions, have in common to record variations of brain activity in time. I’ll first briefly describe those different techniques and the different associated methods of analysis, and then focus on the electrophysiology of epilepsy. Patients with epilepsy present abnormal brain activity, like epileptic spikes and seizures that can be recorded with electroencephalography (EEG). In order to localize the region of the brain that produces this abnormal activity, EEG from the patients is recorded continuously for 2 to 3 weeks. Usual clinical practice involves a neurologist reviewing visually the signal in order to determine the spatial localization and the temporal dynamics of the epileptic activity. While several studies tried to develop an automatic and reliable detection / characterization of the epileptic events in time and space, no fully non supervised methods are commonly used by the neurologists. An efficient time series analysis could be of great interest to speed up the signal analysis and in turn to increase the number of patients handled.

Hands on session for GW data analysis

• Vivien Raymond
• Yann Bouffanais

Room: Room Turing You can find all the hands-on information in the text document.

Musical time series and artificial creative intelligence

• Philippe Esling (IRCAM, Pierre and Marie Curie University)

Room: Room Turing (2)

Hands-on session on visualization and human-computer interaction methods: Designing interactive visualization techniques for large data series

• Anna Gogolou (Inria/Univ.Paris-Saclay/Paris-Descartes Univ)
• Anastasia Bezerianos (Paris-Sud Univ/CNRS/Inria)
• Fanis Tsandilas (Inria/Paris-Sud Univ/CNRS/Paris-Saclay Univ)

Room: Room Turing (2)

Computer science: Privacy-Preserving Use of Individual Smart Metering Data for Customer Services

• Georges Hebrail (EDF)

Room: Room Turing (2)

Discussion: Major Gains and Challenges in Interdisciplinary Time Series Analysis Room: Room Turing (2)