Loosely-Structured Data; Knowledge Graphs; Best-Effort Integration
Smirnova Alisa, Audiffren Julien, Cudre-Mauroux Philippe (2018), Distant Supervision from Knowledge Graphs, in Sherif Sakr and Albert Zomaya (ed.), Springer, Berlin, 1-7.
Rosso Paolo, Yang Dingqi, Cudre-Mauroux Philippe (2018), Knowledge Graph Embeddings, in Sherif Sakr and Albert Zomaya (ed.), Springer, Berlin, 1-7.
Bhardwaj Akansha, Mercier Dominik, Dengel Andreas, Ahmed Sheraz (2017), Deep Learning for Image Based Bibliographic Data Extraction, in ICONIP
, Springer, Berlin.
SmirnovaAlisa, AudiffrenJulien, Cudre-MaurouxPhilippe, APCNN: Tackling Class Imbalance in Relation Extraction through Aggregated Piecewise Convolutional Neural Networks, in SDS 2019
, IEEE, -.
CuccuGiuseppe, TogeliusJulian, Cudre-MaurouxPhilippe, Playing Atari with Six Neurons., in AAMAS 2019
, MontrealAAMAS, Montreal.
Loosely-structured data, which exhibit some degree of structure but whose schemas are unknown, are prominent in Big Data. Yet, they are typically neither properly cataloged nor integrated, leading to absurd processes in which Data Scientists manually have to browse, select, and massage the data. Instead, this proposal proposes an overhaul of integration techniques for loosely-structured information in order to match the volume, velocity and variety of such data. The scientific contribution of this project is divided into two distinct though highly interweaved endeavors: i) the creation of new information extraction and semantic lifting approaches to probabilistically interconnect loosely-structured content from Big Data repositories through incrementally-updated knowledge graphs, and ii) the design of new logical abstractions responsible for crisply exposing the resulting integrated information to Data Scientists through higher-level interfaces.