Stephane Hess
Since November 2021, Stephane Hess (1978) has been appointed as a part-time professor of Decision Modelling, Artificial Intelligence, and mobility research at the Faculty of Technology, Policy, and Management (TPM) for five years. Since 2012, he has been a professor in Leeds and director of the Choice Modelling Centre there. Additionally, he has honorary position at universities of Sydney and Cape Town.
For his work in choice modelling, including its applications in mobility (travel behaviour), energy and the environment, and health, Hess has already been awarded a European Research Council (ERC) grant three times. Furthermore, he is the founder and editor-in-chief of the Journal of Choice Modelling, founder and chair of the steering committee of the International Choice Modelling Conference, and a member of the editorial board of two leading mobility journals. As a consultant, he is active in the industry and government sectors, particularly in transportation, health, and the environment. Along with his colleague David Palma, he developed choice modelling software that is used worldwide.
‘Advanced choice models require high-quality data,’ says Hess. ‘In Delft, among other places, I am pioneering the exploration and exploitation of new data sources, with groundbreaking applications utilising big data.’
Application in new fields
Hess's research involves analysing human decision-making and predicting choice behaviour using advanced mathematical models. A significant focus of his work is on people's travel behaviour. ‘Although most of my applied work is used in the mobility sector,’ says Hess, ‘I am also very active in healthcare, the environment, and economics. Together with my research team at the Choice Modelling Centre in Leeds, we are trying to bridge the gap between different disciplines and apply choice modelling in new fields.’
Powerful combination of disciplines
Hess's newest research line involves integrating techniques from econometrics, mathematical psychology, and artificial intelligence (machine learning). ‘Together with Sander van Cranenburgh from the CiTy AI Lab at TU Delft, I am developing new mathematical modelling approaches that combine choice modelling, machine learning, and mathematical psychology,’ says Hess.
‘These new models have the potential to better explain human decision-making and choice behaviour, and crucially, to make more robust predictions. This is achieved by combining the strengths of different disciplines.’
Software used worldwide
Many years of research and development have resulted in Apollo, a choice modelling software package developed by Hess and his colleague David Palma from the Choice Modelling Centre in Leeds. This software combines the best of existing computation and estimation models and is therefore widely applicable. It is also user-friendly without requiring extensive econometric knowledge. Moreover, it enables more experienced users to develop their own model structures. This freely available software is now used by thousands of professionals worldwide.
‘I am proud of four things,’ says Stephane Hess about the impact of his work. ‘I have pushed the boundaries of science and developed new methods. My research on how travellers 'value' travel time contributes to the justification of transport policy; policymakers ask me to contribute to this. Academically, I have established academic journals and conferences and developed free software that is used worldwide by colleagues. And on the educational front, I help guide the next generation.’
Stronger position for TU Delft
As an 'icon' in the field of decision-making modelling and mobility research, Hess's appointment boosts one of TU Delft's key focus areas: the use of quantitative, formal models and techniques for the analysis and support of decision-making. Its application to travel and new forms of mobility, such as self-driving cars, makes a stimulating contribution to the position of the Department of Transport and Logistics (TLO). TU Delft thus becomes part of the global network of top institutions conducting similar research: from Zurich and the UK to Asia, the US, and Africa.
Helping the next generation
‘One of my achievements that I am perhaps most proud of is my success as a PhD advisor,’ describes Hess his impact on education. In addition to giving various seminars on modelling choices, he supervises several PhD students. He is of great value as a scout and mentor to new generations of academic talent. ‘The students under my supervision achieve excellent publication records with their work,’ acknowledges Hess. ‘Moreover, they regularly receive awards. Two of my students have even received the Eric Pas Dissertation Prize, a prestigious international recognition for Travel Behavior Research.’
Better models for a better future
Hess's work aligns with TU Delft's well-known societal excellence themes, such as the digital society and urbanisation/mobility. Given the developments and transitions in which we find ourselves – also in the field of mobility – a new generation of AI-based decision models is needed to develop new mobility models, for example. Hess: ‘With increasingly better models for human decision-making, we can help policymakers and the industry better predict future needs, provide the right services and products in a timely manner. And we can help them understand how to steer that need.’ And he concludes: ‘Ultimately, better mathematical models of human decision-making are also crucial for the safe implementation of Artificial Intelligence.’
Hess's Personal Passion Pride
‘For me, the student is always in the lead. I can help find a good topic or provide direction. But even though it's a fine line, I always try to let the students lead. Even when we write papers together. Half of my papers are written in collaboration with my students. The better the student, the greater my involvement. But I always sit in the back seat.’
‘TU Delft has a unique mix of researchers’
With the investment from the TU Excellence Fund, the appointment of a postdoc was made possible, as well as the establishment of an extensive data collection. ‘TU Delft has one of the largest transport research groups in the world and an established reputation at the intersection of Decision Modelling, Artificial Intelligence, and mobility research,’ says Hess. ‘As a result, TU Delft has the potential to reach the top of the world and take leadership in this research. Moreover, TU Delft has a unique mix of researchers in the various research areas where I am active. By giving collaboration a more formal basis, we can enter into many impactful collaborations in the coming years.’
Cassandra versus Apollo
The choice modelling software package developed by Hess is called Apollo, named after a god from Greek mythology. ‘The explanation behind that might be fun,’ laughs Hess. ‘Because we couldn't come up with a good acronym, we consulted Greek mythology. Cassandra would have been a logical choice. This goddess had prophetic powers but suffered under the curse that no one wanted to listen to her. That seems a bit like our work: developing predictive models for choice behaviour and trying to sell these ideas to policymakers. Unfortunately, the name Cassandra was already in use. So, we ended up with Apollo, the Greek god of prophecy who passed on this gift to Cassandra.’ (see also www.ApolloChoiceModelling.com).