[DL] Post-doctoral position : Representing and enriching events in knowledge graphs

[Nathalie Aussenac-Gilles]

  Post-doctoral position


  Representing and enriching events in knowledge graphs

Keywords: machine learning on graphs, knowledge graph, event prediction


    Context: XP-event project (2021-2024)

An event is defined as “anything that happens, anything that fits over 
time”: meetings, phone calls, purchases, but also business buyouts, 
change of management, health crises, etc. The events are shared at 
through various communication channels that can be private (internal 
documentation, emails, Slack, Teams, phone, etc.) or public (press, 
Twitter, Facebook, etc.). Knowledge of these events is essential for 
humans to make decisions which themselves will have an impact on future 
events. Many innovative applications can benefit or even emerge from a 
technology capable of extracting events from various sources, 
representing them, aggregating them and exploiting them to predict 
future events. We can for example cite: anticipating demand for sanitary 
products, the supervision of cultural, advertising or festive events, 
but also the study of competition, the study of commercial markets, etc. .

One of the main obstacles to the deployment of these applications is the 
excessively high cost of their development when it is carried out on an 
ad hoc basis by competing players. The XP-Event project proposes to 
respond to this difficulty by setting up a common base for all the 
applications organized around the notion of event. This project is led 
by a consortium naturally formed by two companies (GeoTrend and Emvista) 
and a research team from the IRIT laboratory sharing this vision and 
each having significant scientific and technological heritage in the field.


    Position Description

The candidate will contribute to the tasks in which IRIT is involved, 
and will be more particularly in charge of realizing and implementing 
the proposed solutions. The first task concerns the representation of 
event graphs. The first task will be to define an adapted ontology and a 
process allowing to exploit it to access or represent in RDF the graphs 
of the industrial partners of the project, which are graphs of quite 
different nature. The second task aims at defining a process to evaluate 
the quality of the event graphs. Evaluation will be based on the 
ontology structure as well as on reasoning from the knowledge graph. The 
third task concerns the enrichment of these graphs. Two types of 
approaches will be implemented in the project, and for each of them 
research will be needed to advance the state of the art. The first 
approach consists in extracting information from texts. Each of the 
industrial partners already has its own processing chain that it will 
improve and unify. The second approach consists in exploiting the 
current state of a graph but also the structure of an event in the 
ontology to suggest the addition of new nodes or new relations to the 
graph. This approach will be implemented through learning algorithms 
from graph.


    Requirements for this position

Applicants are required to have a PhD in computer science, and strong 
background ideally in two areas of artificial intelligence: semantic web 
technologies (ontology engineering, linked data management and querying, 
SPARQL, SHACL, RuleML, ...), and machine learning from graphs and vector 
representations, recursive neural networks, etc. Good programming skills 
(Python, OWL API) and experience in participating in collaborative 
projects is required. In addition, the candidate must have a taste for 
innovation, and the ability to dialogue and collaborate with industrial 
partners. Experience in managing graph warehouses (Virtuoso, Strabon, 
Neo4j...) is desired. Fluency in written / spoken English is required 
too. Fluency in French language will be a plus.


    Work environment

Location : Institut de Recherche en informatique de Toulouse (IRIT) - 
UPS, 118 Route de Narbonne F-31062 Toulouse Cedex - France and  et UT2J, 
5 allées Antonio Machado 31300 Toulouse

Duration: 24 months – starting as soon as possible (at best on May 1st, 
2022)

Host team: MELODI 
https://www.irit.fr/en/departement/dep-artificial-intelligence/melodi-team/ 
<https://www.irit.fr/en/departement/dep-artificial-intelligence/melodi-team/>

The candidate will work with four academic researchers from MELODI (F. 
Benamara, Ph. Muller, N. Aussenac-Gilles and N. Hernandez). He will 
collaborate with the partner companies in the project, namely Geotrend, 
located in Toulouse, and Emvista, located in Montpellier.

Income: between 2300 and 2800 euros before taxes (brut) monthly 
according to past experience


    How to apply?

Applicants should send their application files before March 30th 2022 to 
the contact persons listed here below. Application files should contain 
at least a full Curriculum Vitae including a complete list of 
publications, a cover letter indicating their research interests, 
achievements to date and vision for the future, as well as either 
support letters or the name of 2 persons that have worked with them.

Applicants should contact: N. Aussenac-Gilles <aussenac at irit.fr 
<mailto:aussenac at irit.fr>> and N. Hernandez <hernande at irit.fr> 
<mailto:hernande at irit.fr>

-- 
Nathalie Aussenac-Gilles

IRIT - Equipe MELODI
Universite Paul Sabatier
118, route de Narbonne			
31062 TOULOUSE Cedex 9 		

http://www.irit.fr/~Nathalie.Aussenac-Gilles   	aussenac at irit.fr
Tel : +33 5 61 55 82 93 		Fax : +33 5 61 55 62 58
-----------------------------------------------------------------------
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.zih.tu-dresden.de/pipermail/dl/attachments/20220308/9f3e79e4/attachment.htm>
-------------- next part --------------
------------------------------------------------------------------------
Message émis via la liste d'information en Ingénierie des Connaissances
<info-ic at listes.irisa.fr>

Renseignements, inscription :  https://sympa.inria.fr/sympa/info/info-ic
Désinscription : https://sympa.inria.fr/sympa/signoff/info-ic
Archives :    https://sympa.inria.fr/sympa/arc/info-ic
Dépots documents : https://sympa.inria.fr/sympa/d_read/info-ic/
Responsables : Jerome.Nobecourt arrobe univ-paris13.fr, Jean.Charlet  arrobe crc.jussieu.fr
------------------------------------------------------------------------