Four Post doc positions with KRR and CLTL at Vrije Universiteit Amsterdam

The next generation of search technology should be ubiquitous, accurate, flexible, dynamic, intelligent, humane and safe. The DREAMS lab (Dialogues, REAsoning and Multilinguality for Search)will investigate such search systems, focusing on three main approaches: semantic, multilingual and conversational search. The project is a collaboration between the Computational Lexicology & Terminology Lab (CLTL) and the Knowledge Representation and Reasoning (KRR) group, both at the VU, the Information and Language Processing Systems at the UVA and the Huawei Consumer Business Group, a world leading provider of cloud services for customers. At full strength, the DREAMS lab will employ up to 9 people (PostDocs and PhD students).

To start the project the two VU groups (KRR and CLTL) are looking for 4 Postdocs (2 year appointment, with the possibility to extend to 4 years) to work on the following two subprojects:

Your duties

Knowledge-based Text Classification with Explanation: In order to ensure users are not exposed to offensive or otherwise inappropriate online texts the ability to classify content into different categories is necessary. This research aims to provide technical solutions through the use of state-of-the-art Artificial Intelligence (AI) and Machine Learning (ML) approaches to automatically classify text content in different languages. Because current state of the art AI/ML classification methods have little understanding of the reasons why they classify inputs into different classes, this research will also address the challenge of explainability, which could be, e.g., on the basis of large existing (or to be constructed) knowledge graphs.

Multilingual Single Model Information Extraction and Knowledge reasoning System: The goal of this project is to develop an automated multilingual information extraction tool to create high quality knowledge graphs. A critical step in creating such knowledge graphs is the ability for high quality information extraction from text, including named entity Recognition, named entity linking and disambiguation, coreference resolution, temporal information extraction, relation extraction and event extraction.The second goal of this project is to research reasoning over KGs. The research is expected to advance state-of-the-art reasoning and machine learning methods such as distributed embedding based reasoning, neural network based reasoning or hybrid reasoning.

In both subprojects, we will target content in 5 different languages with the ambition to expand this to all European languages.