2-year Postdoc position in Change is Key!
We seek a postdoc for a 2-year position at the University of Gothenburg for semantic change detection across multiple time points.
Applications must be received by: 2022-11-15
Språkbanken Text, the Swedish language bank, (SBX), a nationally and internationally acknowledged unit at the Department of Swedish, Multilinguality, Language Technology, University of Gothenburg is announcing one or more postdoc positions focusing on: identification and analysis of lexical semantic change using computational models applied to diachronic texts. Our languages change over time. As a consequence, words may look the same, but have different meanings at different points in time, a phenomenon called lexical semantic change (LSC). If we do not know which words have changed their meaning and in what ways, we run into problems when interpreting older texts (or more recent texts from different contexts). To facilitate interpretation, search, and analysis of old texts we build computational methods for automatic detection and characterization of LSC from large amounts of text. Our outputs would be directly applicable to the lexicographic R&D unit in our department that compiles the Swedish Academy dictionaries, as well as to researchers from historical linguistics, analytical sociology, gender studies, conceptual history, literary studies, and other research disciplines in the humanities and social sciences that include textual analysis as a central methodological component.
The research is to be conducted in two research projects with related themes where language technology is developed for historical linguistics. More information about the projects is available at https://languagechange.org/ and https://www.changeiskey.org/. Both projects and programs participate in the same context and have joint meetings etc. and contribute to an active and lively research environment for this exciting and growing research field around LSC. Here is an opportunity to contribute to research within LSC but also to social sciences and humanities through our active collaboration within the program with researchers in historical linguistics, analytical sociology, gender studies, history of ideas, and literary studies. A smaller part of the employment, a maximum of 20%, may include tasks other than research, which are relevant to the program as well as pedagogical development work for tutorials and summer courses, alternatively teaching in language technology.
This postdoc position will offer the opportunity to strengthen and enhance scientific skills by conducting research with focus on computational modeling of semantic change together with the researchers at SBX, as well as international researchers from KULeuven (Belgium), IMS Stuttgart (Germany), and Queen Mary University of London (UK). A 3-to-6 month research stay at one of our partner institutions is possible.
The successful applicant will conduct research in NLP, and will develop methods for modeling lexical semantic change for historical texts, with a particular focus on multiple-time point change detection. In particular, you will use deep learning language models and general machine learning methods to develop models for semantic change detection and analysis using large collections of (historical) texts across hundreds of points in time. It is also encouraged to pursue these targets with other methods (e.g., graph theoretical perspectives, modeling of complex (social) networks) You will evaluate your results on English and Swedish document collections, on evaluation sets such as SemEval tasks for word sense induction, and unsupervised semantic change detection, and other existing change testsets. The successful applicant will also help identify and develop relevant and new testsets. If time permits, the postdoc will use the developed models to investigate qualitative hypotheses posted by researchers from history, linguistics, and social sciences. The research tasks will be defined together with the project/program leader to correspond to the project goals, and conducted in an active research environment where in-depth knowledge of methods and evaluation for lexical semantic change exist.
We collaborate by regular meetings in which you will take an active part in the discussions and present results. The postdoc work should contribute to scientific progress which is presented in jointly published journal papers and national and international conference papers. The work should be done independently where the postdoctoral fellow can take responsibility and follow schedules. Together, we will define the direction of the research and it is important to be able to pursue own research tracks while at the same time being able to contribute to the research of others.
The holder of the position is expected to carry out the vast majority of the research in Gothenburg and to actively participate in the department´s workshops, seminars, and conferences, where the main language is English.
To be eligible as a postdoctoral fellow one must have completed a doctoral degree or who has a foreign degree that is deemed to correspond to a completed doctoral degree. This eligibility requirement must be met at the latest at the time when the employment decision is made. The doctoral thesis should be in a relevant area according to the specific position stated here (for example, language technology/natural language processing, computer science, complex systems, mathematics, linguistics with experience in computational methods, or informational sciences). Since the postdoctoral researcher position is designed to give new PhDs the opportunity to strengthen and develop their scholarly proficiency, priority will be given to those whose doctoral degree was granted within 3 years of the application closing date. Due to parental leave, sick leave or military duty, this period can be extended.
Eligibility criteria are described in the Gothenburg University’s Appointments procedure for teachers:
Most important in the assessment of the applicants is their scientific skill, where particular weight will be given to the applicant’s doctoral thesis and other publications. Priority will be given to applicants whose previous work strengthens the research group and the projects in question. Excellent scientific communication skills in English, a high degree of independence and critical thinking are also expected. The applicant must have documented experience and good ability to conduct own research with relevance for the two research projects, of working both independently and in groups, and regularly publish in relevant high-ranking conferences such as ACL, EACL, EMNLP, NeurIPS.
Secondly, experience and ability with modern machine learning frameworks (for example PyTorch or Tensorflow/Keras) as well as large-scale methods for modeling text in Python are strongly meritorious. Experience with LSC detection and/or evaluation methods and having worked with historical or modern, large-scale texts (including experience dealing with noisy texts with OCR errors or social media) is a plus.
Documented experience and ability to work with large-scale text modeling methods (e.g. Word2Vec, FastText, BERT, ELMo), as well as having used such methods for investigations of large-scale text data for information extraction (e.g. information extraction, sentiment analysis, argument extraction) is also meritorious. In addition, very good knowledge of English is a requirement.
For the position, the candidate will be selected who, after an overall assessment of skill in the subject area of the position, is judged to have the best conditions to carry out the tasks included in the position, cooperate with other personnel and contribute to a positive development of the environment.
The recruitment process will consist of an interview (most likely remotely over video link) during which we will discuss how your research and experiences can contribute to computational modeling for semantic change (see https://langsci-press.org/catalog/view/303/3027/2374-1 for a survey on the topic), and a work sample in the form of a programming task or other problem solving task related to the call.
Type of employment: Fixed-term employment, 2 years Extent: 100% Location: Department of Swedish, Multilinguality, language technology First day of employment: According to the agreement
Contact information for the post
If you have any questions about the position, please contact:
Nina Tahmasebi, associate professor. Email: Nina.Tahmasebi@gu.se
Union representatives at the University of Gothenburg can be found here: https://www.gu.se/om-universitetet/jobba-hos-oss/hjalp-for-sokande
In order to apply for a position at the University of Gothenburg, you have to register an account in our online recruitment system. It is the responsibility of the applicant to ensure that the application is complete in accordance with the instructions in the job advertisement, and that it is submitted before the deadline. The selection of candidates is made on the basis of the qualifications registered in the application. If you have any problems or questions, regarding the recruitment system, please contact email@example.com
The application must include:
- curriculum vitae (CV) of both scientific and pedagogical work
- list of degrees and previous employments
- a full list of scientific papers (Attach phd thesis and possibly other published writings, at most 7 works to the application.)
- other documents you wish to refer to track and that can have significant impact on the assessment
- a copy of the doctoral degree
- a brief statement (maximum 1 page) about how your previous research and experiences relate to the project and program.
- Contact details for two references
Applications must be received by: 2022-11-15
The University works actively to achieve a working environment with equal conditions, and values the qualities that diversity brings to its operations.
Salaries are set individually at the University.
In accordance with the National Archives of Sweden’s regulations, the University must archive application documents for two years after the appointment is filled. If you request that your documents are returned, they will be returned to you once the two years have passed. Otherwise, they will be destroyed.
In connection to this recruitment, we have already decided which recruitment channels we should use. We therefore decline further contact with vendors, recruitment and staffing companies.