PhD on Effective Human-AI collaboration via explanations and interactive AI

Updated: 3 months ago
Deadline: 26 Jun 2021

The Human-Technology Interaction Group (HTI) of the School of Innovation Sciences, in collaboration with the Eindhoven Artificial Intelligence Systems Institute (EAISI), is looking for a PhD-student on effective human-AI collaboration via explanations and interactive AI.

Are you interested in improving the way in which humans interact with AI-systems in the domains of health and education? Are you interested in studying how explanations and interactive AI, tailored to the user and its context, could result in optimal human-AI collaboration and thereby improving human functioning in the workplace?

The project
Artificial Intelligence has an increasing impact on the ways in which humans work, live, and learn. In this project we focus on AI systems providing recommendations or suggestions to improve human functioning in the domains of education and health. For example, systems that can help surgeons to identify malignant cells from pathology slides or automatic feedback systems that aid students in their learning. However, the full potential of these systems is often not obtained due to the fact that these systems are not always fully understood or because they are not fully adapted to the user and the (complex) context.

To reach the full potential of these AI systems, it is especially important to have an optimal human-AI collaboration. For this, the AI needs to go beyond merely showing the recommendation. Therefore, this project aims to determine how effective explanations of AI and interactive machine learning can be used to improve human-AI collaboration. The use of explanations and interactivity of AI in health and education tailored to the user's needs and context, has been very limited, with some rare exceptions. Here, we argue that effective explanations of AI are not necessarily synonymous to explaining the underlying logic of the model. Neither might providing full interactivity effectively aid the user. Carefully considering the users' needs in combination with the capabilities of the AI system lies at the heart of this project.

In this project, the human-AI collaboration will be evaluated within two different domains: health and education. By applying this project in two different contexts (health and education), we aim to determine which aspects of explanations and interactive machine learning are more domain-specific and which aspects generalize to other domains as well.

The research groups
The Department of Industrial Engineering & Innovation Sciences (IE&IS) is a leading school in the area of industrial engineering and management science as well as innovation sciences. IE&IS has several BSc and MSc programs housed in two separate schools: Innovation Sciences (IS) and Industrial Engineering (IE). The School of Innovation Sciences focuses on the realization and impact of technological innovations at the individual, organizational, and societal level. The Human-Technology Interaction (HTI) group is part of the school of IS. It concentrates expertise in social sciences - particularly psychology - and engineering, studying technology in the service of human wellbeing within the broader context of a socially and ecologically sustainable society.

The Eindhoven AI Systems Institute (EAISI) combines all TU/e Artificial Intelligence activities. Top researchers from various research groups work together to create new and exciting AI methodologies and applications with a direct impact on the real world. TU/e has been active in the field of AI for many years, which gives the new institute an excellent starting position to build upon.

View or Apply

Similar Positions