Knowledge Graphs, Information Extraction

Department
Date
11-2121
Position
PhD Position – Thesis offer M/F (Reference: DS_PL_PhD_kFLOW_Nov21)

Résumé

History is composed of a continuous flow of events. Each of them can impact subsequent events and contribute to the evolution of human knowledge. Knowledge Graphs (KG) try to encode the information about facts and events, often falling short when representing the temporal evolution of this knowledge and tracking cause-effect flows..

This PhD thesis aims to investigate strategies for extracting, enriching and exploiting event knowledge, applying and developing techniques in the fields of the semantic web, knowledge graphs, information extraction, and deep learning. In particular, the candidate will:

  • Study and develop NLP techniques for the detection of events in the text. In particular, the developed techniques will aim to identify not only the events themselves and the connected entities (agents, time, places, etc.) but also the relations between events. The studied approaches may include bi-directional LSTM, attention-based encoders and word embeddings (BERT). In successive iterations, the techniques will be fine-tuned using the extracted information.
  • Use the extracted information to populate a Knowledge Graph of interconnected events, according to the Semantic Web principles. The KG will be interconnected to external knowledge bases, such as Wikidata, DBpedia, Geonames, EventKG.
  • Applying the knowledge obtained from previous tasks to infer missing relationships and detect invalid ones. The final goal is to expand and correct the KG, taking into account the explainability of the results. The developed techniques will be finally applied in fact-checking scenarios.

This PhD position is fully funded as part of the ANR-JCJC Knowledge Flow (kFLOW) project, which aims to propose strategies for representing, extracting, predicting and using the information about event relationships and knowledge evolution.

Requirements

  • M.Sc.(Eng.) in the area of Computer Science / Computer Engineering or equivalent
  • Good programming skills (preferred Python)
  • Fluency in English (written and spoken)
  • Curiosity and ability to work in a team environment
  • Nice to have: previous experience with Machine Learning and/or Deep Learning

Application
The application must include:

  • Detailed curriculum
  • Transcript of courses taken at graduate and undergraduate levels and their grades
  • Motivation letter of two pages
  • At least 2 recommendation letters

Applications should be submitted by e-mail to pasquale.lisena@eurecom.fr with cc secretariat@eurecom.fr with the reference :  DS_PL_PhD_kFLOW_Nov21

Important Dates
Deadline to apply: 10 December (screening will start immediately)

Start date: January 2022
Duration: duration of the thesis

 

 

 

 

 

 

More info