Workshop

From Hype to Action: Applying Generative AI in Danish Research Practices

DIREC and the Danish Council for Research and Innovation Policy (DFIR) invite to a workshop and a debate on potentials, risks, and barriers for the use of genAI in Danish research and innovation.

The program for the workshop and debate meeting will be updated continuously.

Program

12:30-13:00: Lunch

13:00-14:30: Part 1: Workshop on the application of genAI in research

Three talks on how generative AI has been applied in various research fields, followed by a discussion on how it has influenced research practices.

AI in the humanities: A tool and an object
by Professor Mads Rosendahl Thomsen

The humanities have usually been on the outskirts of research in and with artificial intelligence. With the arrival in 2022 of generally available and broadly used generative AI, not least technologies for creating and analyzing text, this has changed dramatically. I sketch how generative AI influences research, ranging from solid advances in text mining to exploratory approaches, before turning to how the use of generative AI in society is a field with a rapid ascending importance as it affects not least text culture.

Mads Rosendahl Thomsen is Professor of Comparative Literature, Aarhus University. He has worked for the past decade with computational methods in the humanities, both in his research, e.g., the VELUX Fonden supported Fabula-NET project, and as a coordinator of several competency development programs. On April 1, he will lead the new center of excellence TEXT: Center for Contemporary Cultures of Text, supported by the Danish National Research Foundation.

A brief overview of AI trends and technologies in scientific research
by Entrepreneur and Tech Product Lead Mads Rydahl

In this talk, Mads Rydahl will briefly overview how AI has shaped research over the last 15 years. In his previous company, Mads designed recommendation engines used by, for example, Nature.com to suggest related publications. In his new endeavor, he is building a generative AI solution to support the next generation of scientific research. The talk will highlight the development of AI and some of the key challenges for the use of AI in scientific research.

Mads Rydahl is Tech Product Lead at Proemial. He is a seasoned technology and product leader with over two decades of experience driving innovation in the tech industry. He has been recognized with numerous international awards and featured in several publications. As the former Head of Product Design at Siri.com, a startup acquired by Apple in 2010, Mads played a pivotal role in shaping groundbreaking technologies. He is also the first-named inventor on multiple foundational patents in search technology and recommender systems.

Timothy Jenkins, Associate Professor, Department of Biotechnology and Biomedicine, Center for Antibody Technologies, DTU 

14:30-15:00: Break

15:00-15:15:  Part 2: Welcome to debate meeting and presentation of report

15:15-15:30:  Using Generative Artificial Intelligence (GenAI) across different Research Phases – Cases, Potential and Risks
by professor Mads P. Sørensen, Aarhus University

15:30-16:30: Panel debate

Participants are invited to discuss how Danish research environments can move from hype to action in the application of generative artificial intelligence. A particular focus is on stimulating a widespread uptake and application of genAI tools.

Head of Research Ulla Appel Haahr, Business Academy Aarhus, School of Applied Sciences
Chair Janne Gleerup, Danish Association of Masters and PhDs
Dean Anne-Mette Hvas, Health, Aarhus University
Department Head Kaj Grønbæk, Department of Computer Science, Aarhus University
CFO Michael Friis Lindinger, Innovation Fund Denmark

PhD Defence by Manxi Lin

Knowledge-grounded Explainable Medical Image Analysis for Fetal Ultrasound

Abstract

In fetal ultrasound image analysis, neural networks should do more than simply excel at tasks – explanations for the model decisions are also considered important since medical decisions can be high-stakes. This thesis focuses on enhancing model explainability by incorporating knowledge, both human and model-derived, into neural networks.

We study knowledge-grounded explainable artificial intelligence. Our study focuses on two tasks: (1) We create models that are grounded in human prior knowledge, allowing them to “think” like clinicians. The models provide clinician-centered explanations that are useful to the users. (2) We combine human knowledge with insights derived from large-scale pre-trained models to construct interpretable models. The model knowledge enriches the explanations.

We validated our methods across various applications in fetal ultrasound analysis and conducted additional experiments on prostate cancer detection in magnetic resonance imaging (MRI). The results demonstrate that our approach effectively facilitates the model explainability as well as the performance in these applications.

Principal supervisor:

Professor Aasa Feragen, DTU Computer

Co-supervisors:

Associate Professor Anders Nymark Christensen, DTU Compute
Professor Martin Grønnebæk Tolsgaard, KU

Examiners:
Professor Andrew King, Kings College London
Senior Lecturer Maria Zuluaga, EURECOM
Associate Professor Dimitrios Papadopoulos, DTU

Chairperson at defence:

Associate Professor Marco Pizzolato, DTU Compute

A copy of the PhD thesis is available for reading at the department.

Everyone is welcome.

Reception will be held in building 324, room 240

From Hype to Action: Applying Generative AI in Danish Research Practices

Program

Knowledge-grounded Explainable Medical Image Analysis for Fetal Ultrasound

Abstract

Contact info

Newsletter signup