Hi! PARIS, Innovation through a Multidisciplinary Approach to Research and Education
Hi! PARIS is a multidisciplinary center similar to other major research centers created by Institut Polytechnique de Paris and HEC Paris. Why is interdisciplinarity so important for AI and data science?
Eric Moulines : AI is much more than algorithms. This is reason for the partnership between IP Paris and HEC Paris. In engineering schools, we have a tendency to work on methodological issues, the development of new algorithms and new theories explaining these algorithms. This partnership opens new opportunities with the implications of AI. IP Paris and HEC Paris are truly complementary. The Center has three main areas of expertise: 1/ The development of new algorithm methods and theories that support them, 2/ business applications for AI, an area in which HEC Paris has already been working and will continue to focus on, 3/ cross-disciplinary issues involving the applications of AI in social problems.
The topics we work on together, in an interdisciplinary manner, are all related to the applications of AI in our societies: AI and health, AI and law, AI for regulation and AI issues for energy transition. In this sense, Hi! PARIS truly offers a reference point in France and distinguishes itself from “3AI” institutes, created as a result of the report by Cédric Villani, which are technology institutes primarily focused on specific scientific issues.
Thierry Foucault : One particularly innovative aspect of the Center is its multidisciplinary approach. This union of a major research institute in engineering in hard sciences and a world-class business school is unique in France and makes sense. The report by Villani is entitled “For a Meaningful Artificial Intelligence”. To make AI meaningful, it is necessary to both understand AI techniques and the broader social implications, for example in economics, law, ethics, etc.
Therefore, researchers from different areas of expertise must work together at the same research center, covering both data science and social sciences (management, economics, law, psychology, etc.) and promoting a multidisciplinary approach to AI issues. This is precisely why the Center was created.
What will the Center’s research areas be ?
Eric Moulines : On the IP Paris side, we will be working on methods for AI and data science, in other words, the development of algorithms which form the basis for new AI methods. We will also be working on statistical learning and emerging trends, areas where we must position ourselves, such as reinforcement learning, deep learning with methodological and theoretical developments, for example with automatic text processing. We will also focus on the large-scale distribution of algorithms, and the federated learning approach.
The HEC Paris side will develop everything related to understanding data science and AI for business, AI and data in finance, marketing and organizations. We will work together, in an interdisciplinary manner, on subjects related to the applications of AI in our societies.
Thierry Foucault : The Center’s main research topics will be established with the researchers affiliated with the Center and its governance. The goal is to bring together existing research from our institutions regarding, on the one hand, the development and use of AI techniques (e.g. applications for marketing, finance or law) and, on the other hand, the analysis of its economic aspects (e.g. impacts on employment, or the value of data for economic stakeholders), social aspects (e.g. the social acceptability of purely automated decisions), legal aspects (e.g. data ownership issues), and ethics. The goal is also to encourage researchers, through calls for tender, to propose research projects on AI and its implications, to support the development of these projects, and to increase the scientific impact of our institutions in the field of AI by recruiting researchers working on AI in the broadest sense.
How will the Center contribute to education and attractiveness for students?
Thierry Foucault : The purpose of Hi! PARIS is to both produce knowledge on AI and pass this knowledge on in order to allow HEC Paris and IP Paris students to use these techniques and understand all the potential risks and limitations. This is crucial because many of them will find themselves in companies in which data will play an increasingly important role. Why and how can data be used to create value while still contributing to the common good? To answer these questions, we must start by understanding AI techniques and areas of application and then control how these techniques can be used for commercial purposes, while maintaining an awareness of the societal and ethical problems linked to certain uses. HEC Paris and IP Paris are already working together for the Data Science in Business Master’s program, which offers technical training during the first year and training on the managerial applications of AI during the second year. The goal is to increase our capacity to provide this type of training at various levels by increasing the number of research professors contributing to these training programs.
Another aspect is the education of researchers through PhD and post-doctoral programs. More and more, researchers in social sciences are using AI techniques for their research. For example, in my field of research, finance, researchers are increasingly using these techniques to test theories on the valuation of financial assets, develop techniques for managing financial risk, and define certain business strategies through the textual analysis of reports these companies provide to controllers or others.
Eric Moulines : We have very high potential in terms of students. But our students must be able to identify us, to see us as major stakeholders in the field. This is one of the key issues for the Center. We have a lot of very good students who go to Stanford and MIT. This perception among students is what must change. We are a major stakeholder, but it is not sufficiently well known. Why? If we look at the potential of researchers, who publish in the best journals in the field, we already form a very significant group. But we are divided in several institutions and there is not enough visibility for this research. There is visibility within institutions, but not enough for our students. We must be capable of attracting the best European and international researchers and the brightest students.