Biography
Dr Jan-Peter Calliess has been a Senior Research Fellow (equivalent to Associate Research Professor in the US system) at the Department of Engineering Science and Oxford-Man Institute of Quantitative Finance since May 2017.
Between November 2014 and April 2017, he was a Research Associate (postdoc) at the Engineering Department at the University of Cambridge. Working at the intersection of machine learning and control under Jan Maciejowski and Carl Rasmussen, he was a member of both the Computational and Biological Learning Lab and the Control Group. While in Cambridge, he was working within the Autonomous and Intelligent Systems Partnership (AISP) and was grateful for having received funding from EPSRC and Schlumberger. He also worked on learning-based computational mechanism design, recieving an award in support of this work from EPSRC-NCSML.
Before coming to Cambridge, Jan-Peter was a DPhil student with the Pattern Analysis and Recognition Group at Oxford, funded by an EPSRC studentship and the ORCHID project. Before that, he received his undergraduate/postgraduate degree in Computer Science from University of Karlsruhe (TH) and had repeated stays at SCS, Carnegie Mellon University between the years of 2006 and 2009.
Research Interests
Jan-Peter's interests include a wide range of topics pertaining to decision-making and computational learning. He has published in various communities related to artificial intelligence, including papers on machine learning, control, signal processing, optimisation and multi-agent systems.
Apart from working on new learning algorithms and their theoretical understanding, he is also currently investigating applications of machine learning to finance, control and dynamical systems.
Service
Jan-Peter served on the programme comittee and as a reviewer for Automatica, IEEE Trans. on Automatic Control, IEEE Robotics and Automation Letters, ICRA, ICML, NIPS, IROS, Multidimensional Signals and Systems, ACC, ECC, CDC, NMPC as well as for the Swiss National Science Foundation (SNSF).
Formal Publications
J. Calliess, A. Papachristodoulou and S. J. Roberts. Bayesian nonparametrics and feedback-linearisation of discretised control-affine systems. To appear in Proc. of the Conf. on Decicion and Control (CDC), 2018.
J. M. Manzano, D. Limon, D. Munoz de la Pena and J. Calliess. Robust Data-Based Model Predictive Contol for Nonlinear Constrained Systems. To appear in NMPC 2018.
A. Edwards, J. Calliess and K. Margellos. Distributed Optimization for Energy Management in Building Networks. To appear in Control 2018.
J. Calliess, S. J. Roberts, C. E. Rasmussen and J. Maciejoswki.Nonlinear Set Membership Methods with Hyperparameter Estimation for Online Learning and Control. Proc. of the ECC, 2018.
A. Blaas, A. Cobb, J. Calliess and S. J. Roberts.Scalable Bounding of Predictive Uncertainty in Regression Problems with SLAC. To appear in 12th Int. Conf. on Scalable Uncertainty Management (SUM), 2018.
J. Calliess Lipschitz Optimisation for Lipschitz Interpolation. Proc. of the ACC, 2017.[LINK]
D. Limon, J. Calliess and J. Maciejowski. Learning-based Nonlinear Model Predictive Control . Proc. of the 2017-IFAC World Congress, 2017.
J. Calliess, N. Korda, G. J. Gordon. A Distributed Mechanism for Multi-Agent Convex Optimisation and Coordination with No-Regret Learners, Workshop on Learning, Inference and Control of Multi-Agent Systems, NIPS, 2016.
J. Calliess. Bayesian Lipschitz Constant Estimation and Quadrature, Workshop on Probabilistic Integration, NIPS, 2015.
J. Calliess, M. Osborne and S. J. Roberts. Conservative collision prediction and avoidance for stochastic trajectories in continuous time and space. Proc. Autonomous Agents and Multi-agent Systems (AAMAS), 2014.
J. Calliess A. Papachristodoulou and S. J. Roberts. Stochastic processes and feedback-linearisation for online identification and Bayesian adaptive control of fully-actuated mechanical systems, WS- Advances in Machine Learning for Sensorimotor Control, NIPS, 2013. (Also submitted to Arxiv)
J. Calliess, M. Osborne and S. J. Roberts. Nonlinear adaptive hybrid control by combining Gaussian process system identification with classical control laws, WS- Novel Methods for Learning and Optimization of Control Policies and Trajectories for Robotics, ICRA, 2013.
J. Calliess and S. J. Roberts. Multi-agent planning with mixed-integer programming and adaptive interaction constraint generation. (Extended Abstract), Symposium on Combinatorial Search (SOCS), 2013.
J. Calliess, M. Osborne and S. J. Roberts. Towards auction-based multi-agent collision-avoidance under continuous stochastic dynamics. Presented at workshop: Markets, Mechanisms, and Multi-Agent Models — Examining the Interaction of Machine Learning and Economics, (ICML 2012).
D. Lyons, J. Calliess and U. Hanebeck. Chance-constrained Model Predictive Control for Multi-Agent Systems. Proc. of the American Control Conference (ACC 2012).
J. Calliess, D. Lyons and U. Hanebeck. Lazy auctions for multi-robot collision avoidance and motion control under uncertainty. LNAI 7068, Springer, 2011.
J. Calliess, M. Mai, S. Pfeiffer. On the Computational Benefit of Tensor Separation for High-Dimensional Discrete Convolutions. Multidimensional Systems and Signal Processing, Springer, 2010.
J. Calliess. On Fixed Convex Combinations of No-Regret Learners. 6th International Conference on Machine Learning and Data Mining in Pattern Recognition. In LNAI 5632, Springer, 2009.
S. Pfeiffer, M. Mai, W. Globcke, J. Calliess. On generalized separation and the speed-up of local operators on multi-dimensional signals. 6th International Workshop on Multidimensional (nD) Systems (NDS ’09). (IEEE-XPLORE).
A. Porbadnigk, M. Wester, J. Calliess, T. Schultz. EEG-based Speech Recognition – Impact of Temporal Effects. International Conference on Bio-inspired Systems and Signal Processing, Biosignals 2009.
J. Calliess and G. J. Gordon. No-regret Learning and a Mechanism for Distributed Multiagent Planning. Proc. of the 7th International Conference on Autonomous Agents and Multiagent Systems (AAMAS), 2008.
Misc.
J. Calliess. Online Optimisation for Online Learning and Control - From No-Regret to Generalised Error Convergence. ICML/IJCAI/AAMAS Workshop on Planning and Learning (PAL), 2018.
J. Calliess. Lipschitz Optimisation for Lipschitz Interpolation. Arxiv preprint, 2017. [PDF-LINK]
J. Calliess. Lazily Adapted Constant Kinky Inference for Nonparametric Regression and Model-Reference Adaptive Control, arXiv:1701.00178, 2016. [PDF-LINK]
J. Calliess. Conservative decision-making and inference in uncertain dynamical systems. DPhil thesis. University of Oxford, 2014.
J. Calliess, M. Osborne and S. J. Roberts. Conservative collision prediction and avoidance for stochastic trajectories in continuous time and space. Arxiv Preprint, 2014. [PDF-LINK] (An improved version can be found in my thesis above).
J. Calliess, M. Osborne and S. J. Roberts. Towards optimization-based multi-agent collision-avoidance under continuous stochastic dynamics. Presented at AAAI-2012, Workshop on Multiagent Pathfinding, Toronto, Canada, 2012.
J. Calliess, D. Lyons and U. Hanebeck. Lazy auctions for multi-robot collision avoidance and motion control under uncertainty. Technical Report. No: PARG-11-01. University of Oxford. 2011. (Extended version of workshop publication above).
J.-P. Calliess, On Fixed Convex Combinations of No-Regret
Learners. Technical Report. Machine Learning Dept., Carnegie Mellon University, 2008.
J.-P. Calliess, and G. J. Gordon. No-Regret Learning and a Mechanism for Distributed Multi-agent Planning. Technical Report. Machine Learning Dept., Carnegie Mellon University, 2008. [PDF-LINK]
J.-P. Calliess. Diplomarbeit. No-regret Learning and Market-based Multiagent Planning. IES, Fakultaet fuer Informatik, Universitaet Karlsruhe. September 2007
