Postdoctoral Researcher for Natural Legal Language Processing

Updated: about 2 years ago
Deadline: 14 Mar 2022

How does counting and computation transform the assumptions, operations and outcomes of the law? That is what you aim to find out as a postdoc. You will delve deeper into the use of natural language processing for legal search and quantified legal prediction. This way, you can combine cutting-edge research into computer science with societally relevant problems.

The institute for Computing and Information Sciences is looking for a postdoctoral researcher in the domain of NLP or information retrieval, who will contribute to the study of natural legal language processing (NLLP) and is capable of developing an internal critique of relevant CS methodologies. You will be part of the CS team working on an ERC Advanced Grant project on the subject of 'computational law', in which a legal team and a CS team collaborate to help lawyers better understand both the possibilities and limitations of NLLP. We expect excellence within the domain of CS as well as a willingness to investigate the assumptions of NLP and their implications for the outcome of NLLP systems. You will be appointed at the Radboud institute for Computing and Information Sciences (iCIS), where your colleagues will be leading experts on information retrieval and NLP (e.g. Martha Larson and Arjen de Vries). You will also have the opportunity to collaborate with NLP experts at Radboud University's Centre for Language and Speech Technology. In addition, iCIS has close ties with Radboud University's interdisciplinary research hub on digitalization and society (the iHub). You will partake in the cross-disciplinary research of the ERC Advanced Grant project on 'Counting as a Human Being in the Era of Computation Law (COHUBICOL)'.

You will inquire into the architecture and foundational issues of NLLP, developing answers to questions such as:

1. As machine learning is claimed to approximate a mathematical target function, how can lawyers learn to test and contest the relevance of the approximation for the legal issues at stake?
2. How do different types of NLLP restrict or enable testing for various types of bias in their output?



Similar Positions