The transregional collaborative research center (TRR) “Constructing Explainability” funded by the German Research Foundation (DFG) at the Universities of Paderborn and Bielefeld, deals with explainable and explanatory artificial intelligence.
The central hypothesis of the TRR is that explanations should be co-constructed by dialogue partners so that humans can make confident and informed decisions when interacting with intelligent systems.
To this end, both universities aim to contribute basic research in the areas of co-constructive explanation processes and their mechanisms, explanation as a socio-technical phenomenon, and computational representations of the dynamics of co-constructive explanation.
They bring together different disciplinary perspectives both at the level of individual projects and the TRR as a whole.
The goal of the second funding phase is to develop more flexible, personalized, proactive explanation systems (sXAI) that take social and situational context into account when adapting their explanatory behavior.
The sub-project C03 Explaining Change focuses on ex-ante explanations for the adaptation of complex, composite AI systems to a new context or a new task.
Prominent examples are fine-tuning, transfer learning or prompting of complex AI systems that are grouped around a basic model, but also the adaptation of more classical hybrid systems.
A special challenge in the project is the suitable representation of the different parameterizations of context, system and, if necessary, multi-criteria objective function as well as the handling of uncertainty in the XAI technologies to be developed.
In addition to mathematical modeling and algorithmic implementation, evaluation is also part of the task.
The position will be filled subject to project approval.
Your Tasks
research activities (95 %): mathematical modeling of XAI technologies for AI systems
algorithmic implementation of machine learning approaches
evaluation of approaches (crowd-sourcing platform)
evaluation, publication and presentation of scientific results cooperation with network partners and participation in TRR activities (5%) Employment is conducive to academic qualification.
We offer
salary according to Remuneration level 13 TV-L fixed-term (limited until - end of project) (§ 2 (1) sentence 1 of the WissZeitVG; in accordance with the provisions of the WissZeitVG and the Agreement on Satisfactory Conditions of Employment, the length of contract may differ in individual cases) fulltime internal and external training opportunitiesvariety of health, consulting and prevention servicesreconcilability of family and workflexible working hoursgood transport connectionsupplementary company pensioncollegial working environmentopen and pleasant working atmosphereexciting, varied tasksmodern work environment with digital processes various offers (canteen, cafeteria, restaurants, Uni-Shop, ATM, etc.) Your Profile We expect
completed scientific university degree (e.
g.
Master's or comparable) in computer science, mathematics or in related fieldsstrong knowledge of machine learning (theory and practice)very good knowledge of mathematics (modeling, optimization, statistics)interest and passion for interdisciplinary research at the interface of machine learning and cognitive sciencevery good programming skills (Python)good knowledge of German and English (oral and written)analytical skills and scientific mindsetindependent, self-reliant and committed way of working Preferred experience and skills
practical experience in research on large language models, in particular open weight LLMsknowledge of XAI Application Procedure
We are looking forward to receiving your application.
To apply, please preferably use our online form via the application button below.
application deadline: Contact Prof.
Dr. Barbara Hammer
+49 521 106-12115
Postal Address Universität Bielefeld
Technische Fakultät
Prof.
Dr. Barbara Hammer
Postfach 10 01 31
33501 Bielefeld