HPD Projects
Still looking for a project for your Honours Programme Bachelor or Honours Programme Master?
Are you interested in doing a Machine Learning-oriented project with one of the ELLIS Delft Faculty Members?
The ELLIS Delft Unit focuses on using machine learning as a key enabling technology to deal with complex tasks, and making intelligent systems adapt to their environment including handling human interaction. Some of the ELLIS Delft Unit staff members have proposed directions in which they would be interested in supervising HPD research projects.
Why choose a project with an ELLIS faculty member?
- Collaboration with a community of ELLIS HPD students (e.g. meetings, workshops, collaborative projects)
- Yearly ELLIS HPD event
- ELLIS members can facilitate connections with other units
Want to apply for a project? Send an e-mail directly to an ELLIS staff member (see sample projects below).
A Few Possible Projects:
| Possible Projects | Contact |
|---|---|
| Deep learning solutions for complex networks. A.k.a. geometric deep learning. With applications to recommender systems. | Elvin Isufi, EEMCS, Intelligent Systems, Multimedia Computing |
| Bayesian RL; Model-based RL; Causal RL; Interactive learning and non-stationarity. | Frans A. Oliehoek, EEMCS, Intelligent Systems, Interactive Intelligence |
| Social signal processing, multisensor fusion, computer vision, pattern recognition and machine learning, and ubiquitous computing. | Hayley Hung, EEMCS, Intelligent Systems, |
| Machine Learning for multi-robot control and/or motion planning. Applications in self-driving cars, drones; mobile manipulators and ridesharing. | Javier Alonso Mora, ME, Delft Center for Systems and Control J.AlonsoMora@tudelft.nl |
| Machine Learning applied to the movement of robot arms. Topics: Robotics, Motor Skill Learning, Reinforcement Learning, Imitation Learning, Deep Learning, Interactive Learning | Jens Kober, ME, Cognitive Robotics, Learning and autonomous control |
| How can we use Machine Learning methods to answer causal questions? Causal models for improving generalization of ML models. | Jesse Krijthe, EEMCS, Intelligent Systems, Pattern Recognition & Bioinformatics |
| Probabilistic Machine Learning, Adversarial Learning, Robust Machine Learning. | Luca Laurenti, ME, Delft Center for Systems and Control |
| Responsible AI, Interactive machine learning, fairness in AI, explainable AI. | Luciano Cavalcante Siebert, EEMCS, Intelligent Systems, Interactive Intelligence |
| Field of sensor AI, combining physics-based models with models learned from data (specifically using Gaussian processes). | Manon Kok, ME, Delft Center for Systems and Control |
| Reinforcement learning for motion control; Vision-based robotics (navigation, grasping); Symbolic regression, incl. NN based methods. | Robert Babuska, ME, Cognitive Robotics, Learning and autonomous control R.Babuska@tudelft.nl |
| Learning models from software data. Machine learning that operate more based on logic than statistics | Sicco Verwer, EEMCS, Intelligent Systems, Cyber Security |
| Anomaly detection, dynamical systems, (stochastic) optimization, decision-making under uncertainty | Peyman Mohajerin Esfahani, ME, Delft Center for Systems and Control |
ELLIS (European Laboratory for Learning and Intelligent Systems) is a European network that promotes research excellence and advances breakthroughs in AI. Focusing on fundamental science, technical innovation and societal impact, the main goal of the network is to support European values in shaping the impact of AI, ensure the highest level of AI research, and keep academia attractive for AI researchers.