Human AI Complementarity in Computer Vision

  • Background

    In the last few years, many AI algorithms have been developed—and successfully applied in multiple domains—that can perform tasks with at least the same accuracy as domain experts. This unprecedented development led to the discussion of whether such AI systems will replace human domain experts in the foreseeable future. For example, in the context of medical diagnostics, the research community is actively investigating whether human experts or the AI should make a medical diagnosis. Recent studies have shown that superior results can be achieved when the predictions of AI systems are not considered separately but combined with the knowledge of domain experts. The idea of this "human-AI complementarity" is to regard the AI not as a competitor to the domain expert but instead as a complementary team member. The aim is to consider the individual strengths of both humans and AI. To enable this type of collaboration it is required to understand which instances are difficult for humans to predict or difficult for the AI. 


    Research Goal

    In this thesis, we want to study how a collaborative system consisting of human experts and an AI algorithm must be designed to leverage the individual strengths to improve the overall team performance. Using available scientific datasets, this thesis' objective is to implement a state-of-the-art classification or semantic segmentation model considering human experts strengths and weaknesses to realise the complementary potential between humans and AI. In a second step, we aim to develop and implement an AI system that distributes new instances between human experts and the AI model based on their individual strengths. Finally, we want to evaluate whether the collaboration between humans and AI can result in superior performance. In detail, you will...

     - apply your theoretical knowledge to a practical use case.
     - become an expert in deep learning.
     - develop, implement, and evaluate computer vision models. 


    We look forward to receiving your application because you...

     - have a solid technical understanding of machine and deep learning techniques. 
     - are proficient in Python programming (e.g., pandas, tensorflow, pytorch, scikit-learn). 
     - work in an independent and accurate way.



    Start: immediately | Duration: 6 Month


    We offer you a challenging research topic, close supervision, and the opportunity to develop practical and theoretical skills. If you are interested, please send your CV, transcript of records, and a brief letter of motivation to