PhD - Automated Machine Learning for Optimal Hybrid Models
- Robert-Bosch-Campus 1, 71272 Renningen, Germany
- Legal Entity: Robert Bosch GmbH
Do you want beneficial technologies being shaped by your ideas? Whether in the areas of mobility solutions, consumer goods, industrial technology or energy and building technology – with us, you will have the chance to improve quality of life all across the globe. Welcome to Bosch.
The Robert Bosch GmbH is looking forward to your application!
Hybrid Modeling (in the sense of physics-guided machine learning) has emerged as a promising field of research in machine learning. It aims at leveraging prior domain knowledge, from e.g. environmental physics, to guide data-driven machine learning towards more informed and explainable predictive models. However, designing and optimizing a hybrid model for a given use-case is usually a laborious and time-consuming task. Each new use-case is best addressed with a certain hybridization paradigm, which in turn comes along with its own hyperparameter optimization challenges.
The problem we address in this PhD is the laborious design, selection, and training of hybrid models for each newly defined use case. We are interested in reducing the manual effort in this process to minimal, starting from the selection of the appropriate hybridization mechanism given a use case, and down to hyperparameter optimization on selected models. To this end, you will work in this project on a novel machine learning automation framework (AutoML) adapted to hybrid modeling use cases.
In addition to the several challenges in standard automation frameworks, "AutoML for Optimal Hybridization" faces new interesting research questions. For instance, hybridization introduces more complexities on the permissible configuration spaces for hyper-parameters. It also entails a careful design of fair evaluation metrics and selection criteria for hybrid models. Additionally, superior performance of selected models needs to be validated not only on attainable real-world measurements but also as to which extent physical constraints have been satisfied.
During your PhD, you will be part of the active research team at the Bosch Center for Artificial Intelligence (BCAI) and part of the publicly funded project "Proper Hybrid Models for Smarter Vehicles (PHyMoS) [https://phymos.de]".
- You will develop and implement novel AutoML approaches for Hybrid Models.
- You will conduct prototypical implementations of your ideas, and benchmark on both synthetic scenarios and real-world use cases.
- Furthermore, you will publish in top-tier machine learning conferences (ICML, NIPS, ICLR, AISTATS etc.) and journals (JMLR, PAMI etc.).
- You will screen literature and build up close contact with the academic community.
- You will be part of the BCAI research team, participate in academic interactions and technical discussions within BCAI research, and have the chance of integrating your developments in real industrial applications.
- Education: you have achieved an excellent Master of Science degree in mathematics, computer science, cybernetics, physics or comparable subject
- Personality and Working Practice: Strong and motivated team player who has the ability to independently pursue research work and likes to work in an interdisciplinary and international team
- Experience and Knowledge: Experience in development and implementation of state-of-the-art machine learning technologies and proven programming skills, in particular Python or Matlab
- Languages: fluent in English (written and spoken)
The final PhD topic is subject to your university. Duration: 3 years
Please submit all relevant documents (incl. curriculum vitae, transcripts, certificates, motivation letter, publications, blog posts and GitHub repos, if available)
Apply now in just 3 minutes!
Are you interested in working from home, or part-time? Please don't hesitate to ask us.
Need support during your application?
Kevin Heiner (Human Resources)
+49 711 811 12223
Need further information about the job?
Karim Barsim (Functional Department)
+49 174 3650742