Post-doc position on grey-box machine learning (closed)

The Data Mining and Machine Learning group (http://dmml.ch/) at the University of Applied Sciences in Geneva has an opening for a full-time post-doc position. The research target is to develop grey-box (hybrid) machine learning methods that combine data-driven models such as deep neural nets and theory-driven, physical and/or causal models [e.g., 1,2,3]. By combining these two regimes of modeling, we expect to improve not only the performance but also the interpretability of the outputs of machine learning. Learning such grey-box models causes a number of technical challenges. For example, empirical risk minimization may result in a meaningless combination of the two models with theory-driven components being ignored and overwritten by deep neural nets, which necessitates constraining or regularizing the model [1,3]. Moreover, combining non-differentiable theory models with neural nets can hardly be done with standard tools of optimization. We will focus on such technical challenges in learning grey-box models and related problems, building upon our recent work on deep grey-box models [3,4].

The position is funded by a 3-year project of the Swiss National Science Foundation (SNSF) in the frame of the Strategic Japanese-Swiss Science and Technology Programme (SJSSTP). The project is for developing interpretable condition monitoring methods for complex engineering systems. The aforementioned grey-box machine learning methods will be utilized as building blocks of a condition monitoring (e.g., anomaly detection) framework that should hold a certain extent of interpretability. The research will be conducted in close collaboration between the DMML group and a Japanese counterpart, the Artificial Intelligence Lab at RCAST, the University of Tokyo (https://sites.google.com/g.ecc.u-tokyo.ac.jp/ailab/top-english), who will mainly work on the implementation of a condition monitoring framework and its deployment on real-world systems.

We seek strongly motivated candidates dedicated to high-quality research. Candidates should have (or be close to obtaining) a PhD in machine learning or related areas and a strong research track record attested by high-quality publications in relevant machine learning venues such as ICML, NeurIPS, ICLR, AISTATS, UAI, etc. The selected candidate is expected to demonstrate a high degree of independence and autonomy, drive their own research, and actively contribute to the scientific development of the group through their knowledge and expertise as well as by proposing and contributing to group activities such as readings, schools, workshops, etc. There is also a possibility to participate in supervising a PhD student working on the same or related projects and in teaching activities.

The DMML group consists of around ten researchers at the PhD and post-doc levels, working in different areas of machine learning, such as generative models, simulation-based inference, and imitation learning. The team collaborates closely with the VIPER group (https://viper.unige.ch) from the Computer Science Department of the University of Geneva headed by Prof. Stephane Marchand-Maillet. In addition, the group is involved in a number of national and international research projects. We offer ample opportunities and support for scientific development, e.g. providing funding for conferences, schools, research visits, exchanges, etc. We strive to provide a research environment in which researchers can focus on their research and allow for space and time to develop solid ideas.  

If interested, please send the following both to Prof. Alexandros Kalousis (alexandros.kalousis@hesge.ch) and Dr. Naoya Takeishi (naoya.takeishi@hesge.ch) with the email subject being “SJSSTP Postdoc Application”.

- academic CV (max 2 pages)
- pointers to the candidate’s two most important publications
- one-page motivation letter explaining why the candidate is suitable for the position and what they can bring to the DMML group
- 1000-word research proposal on the research topic of grey-box machine learning
- contact details of three referees (please do not send reference letters)
- copies of diplomas (PhD, MSc, BSc) and academic transcripts

Deadline for applications: 30 April 2023

In case of any further questions, please contact
alexandros.kalousis@hesge.ch and/or naoya.takeishi@hesge.ch.

[1] Y. Yin et al., Augmenting physical models with deep networks for complex dynamics forecasting, in ICLR 2021.
[2] Z. Qian et al., Integrating expert ODEs into neural ODEs: Pharmacology and disease progression, in NeurIPS 2021.
[3] N. Takeishi and A. Kalousis, Physics-integrated variational autoencoders for robust and interpretable generative modeling, in NeurIPS 2021.
[4] N. Takeishi and A. Kalousis, Deep grey-box modeling with adaptive data-driven models toward trustworthy estimation of theory-driven models, arXiv:2210.13103, to appear in AISTATS 2023.