TY - GEN
AB - Many problems in science and engineering require the efficient numerical
approximation of integrals, a particularly important application being the
numerical solution of initial value problems for differential equations. For
complex systems, an equidistant discretization is often inadvisable, as it
either results in prohibitively large errors or computational effort. To this
end, adaptive schemes have been developed that rely on error estimators based
on Taylor series expansions. While these estimators a) rely on strong
smoothness assumptions and b) may still result in erroneous steps for complex
systems (and thus require step rejection mechanisms), we here propose a
data-driven time stepping scheme based on machine learning, and more
specifically on reinforcement learning (RL) and meta-learning. First, one or
several (in the case of non-smooth or hybrid systems) base learners are trained
using RL. Then, a meta-learner is trained which (depending on the system state)
selects the base learner that appears to be optimal for the current situation.
Several examples including both smooth and non-smooth problems demonstrate the
superior performance of our approach over state-of-the-art numerical schemes.
The code is available under https://github.com/lueckem/quadrature-ML.
AU - Dellnitz, Michael
AU - Hüllermeier, Eyke
AU - Lücke, Marvin
AU - Ober-Blöbaum, Sina
AU - Offen, Christian
AU - Peitz, Sebastian
AU - Pfannschmidt, Karlson
ID - 21600
T2 - arXiv:2104.03562
TI - Efficient time stepping for numerical integration using reinforcement learning
ER -
TY - JOUR
AB - It is a challenging task to identify the objectives on which a certain decision was based, in particular if several, potentially conflicting criteria are equally important and a continuous set of optimal compromise decisions exists. This task can be understood as the inverse problem of multiobjective optimization, where the goal is to find the objective function vector of a given Pareto set. To this end, we present a method to construct the objective function vector of an unconstrained multiobjective optimization problem (MOP) such that the Pareto critical set contains a given set of data points with prescribed KKT multipliers. If such an MOP can not be found, then the method instead produces an MOP whose Pareto critical set is at least close to the data points. The key idea is to consider the objective function vector in the multiobjective KKT conditions as variable and then search for the objectives that minimize the Euclidean norm of the resulting system of equations. By expressing the objectives in a finite-dimensional basis, we transform this problem into a homogeneous, linear system of equations that can be solved efficiently. Potential applications of this approach include the identification of objectives (both from clean and noisy data) and the construction of surrogate models for expensive MOPs.
AU - Gebken, Bennet
AU - Peitz, Sebastian
ID - 16295
JF - Journal of Global Optimization
TI - Inverse multiobjective optimization: Inferring decision criteria from data
VL - 80
ER -
TY - GEN
AB - The Koopman operator has become an essential tool for data-driven approximation of dynamical (control) systems in recent years, e.g., via extended dynamic mode decomposition. Despite its popularity, convergence results and, in particular, error bounds are still quite scarce. In this paper, we derive probabilistic bounds for the approximation error and the prediction error depending on the number of training data points; for both ordinary and stochastic differential equations. Moreover, we extend our analysis to nonlinear control-affine systems using either ergodic trajectories or i.i.d.
samples. Here, we exploit the linearity of the Koopman generator to obtain a bilinear system and, thus, circumvent the curse of dimensionality since we do not autonomize the system by augmenting the state by the control inputs. To the
best of our knowledge, this is the first finite-data error analysis in the stochastic and/or control setting. Finally, we demonstrate the effectiveness of the proposed approach by comparing it with state-of-the-art techniques showing its superiority whenever state and control are coupled.
AU - Nüske, Feliks
AU - Peitz, Sebastian
AU - Philipp, Friedrich
AU - Schaller, Manuel
AU - Worthmann, Karl
ID - 23428
T2 - arXiv:2108.07102
TI - Finite-data error bounds for Koopman-based prediction and control
ER -
TY - GEN
AB - As in almost every other branch of science, the major advances in data
science and machine learning have also resulted in significant improvements
regarding the modeling and simulation of nonlinear dynamical systems. It is
nowadays possible to make accurate medium to long-term predictions of highly
complex systems such as the weather, the dynamics within a nuclear fusion
reactor, of disease models or the stock market in a very efficient manner. In
many cases, predictive methods are advertised to ultimately be useful for
control, as the control of high-dimensional nonlinear systems is an engineering
grand challenge with huge potential in areas such as clean and efficient energy
production, or the development of advanced medical devices. However, the
question of how to use a predictive model for control is often left unanswered
due to the associated challenges, namely a significantly higher system
complexity, the requirement of much larger data sets and an increased and often
problem-specific modeling effort. To solve these issues, we present a universal
framework (which we call QuaSiModO:
Quantization-Simulation-Modeling-Optimization) to transform arbitrary
predictive models into control systems and use them for feedback control. The
advantages of our approach are a linear increase in data requirements with
respect to the control dimension, performance guarantees that rely exclusively
on the accuracy of the predictive model, and only little prior knowledge
requirements in control theory to solve complex control problems. In particular
the latter point is of key importance to enable a large number of researchers
and practitioners to exploit the ever increasing capabilities of predictive
models for control in a straight-forward and systematic fashion.
AU - Peitz, Sebastian
AU - Bieker, Katharina
ID - 21199
T2 - arXiv:2102.04722
TI - On the Universal Transformation of Data-Driven Models to Control Systems
ER -
TY - CONF
AB - The first order optimality conditions of optimal control problems (OCPs) can
be regarded as boundary value problems for Hamiltonian systems. Variational or
symplectic discretisation methods are classically known for their excellent
long term behaviour. As boundary value problems are posed on intervals of
fixed, moderate length, it is not immediately clear whether methods can profit
from structure preservation in this context. When parameters are present,
solutions can undergo bifurcations, for instance, two solutions can merge and
annihilate one another as parameters are varied. We will show that generic
bifurcations of an OCP are preserved under discretisation when the OCP is
either directly discretised to a discrete OCP (direct method) or translated
into a Hamiltonian boundary value problem using first order necessary
conditions of optimality which is then solved using a symplectic integrator
(indirect method). Moreover, certain bifurcations break when a non-symplectic
scheme is used. The general phenomenon is illustrated on the example of a cut
locus of an ellipsoid.
AU - Offen, Christian
AU - Ober-Blöbaum, Sina
ID - 22894
KW - optimal control
KW - catastrophe theory
KW - bifurcations
KW - variational methods
KW - symplectic integrators
TI - Bifurcation preserving discretisations of optimal control problems
ER -
TY - JOUR
AB - In this article, we present an efficient descent method for locally Lipschitz
continuous multiobjective optimization problems (MOPs). The method is realized
by combining a theoretical result regarding the computation of descent
directions for nonsmooth MOPs with a practical method to approximate the
subdifferentials of the objective functions. We show convergence to points
which satisfy a necessary condition for Pareto optimality. Using a set of test
problems, we compare our method to the multiobjective proximal bundle method by
M\"akel\"a. The results indicate that our method is competitive while being
easier to implement. While the number of objective function evaluations is
larger, the overall number of subgradient evaluations is lower. Finally, we
show that our method can be combined with a subdivision algorithm to compute
entire Pareto sets of nonsmooth MOPs.
AU - Gebken, Bennet
AU - Peitz, Sebastian
ID - 16867
JF - Journal of Optimization Theory and Applications
TI - An efficient descent method for locally Lipschitz multiobjective optimization problems
VL - 188
ER -
TY - JOUR
AU - Goelz, Christian
AU - Mora, Karin
AU - Stroehlein, Julia Kristin
AU - Haase, Franziska Katharina
AU - Dellnitz, Michael
AU - Reinsberger, Claus
AU - Vieluf, Solveig
ID - 21195
JF - Cognitive Neurodynamics
TI - Electrophysiological signatures of dedifferentiation differ between fit and less fit older adults
ER -
TY - JOUR
AU - Klus, Stefan
AU - Gelß, Patrick
AU - Nüske, Feliks
AU - Noé, Frank
ID - 24170
JF - Machine Learning: Science and Technology
SN - 2632-2153
TI - Symmetric and antisymmetric kernels for machine learning problems in quantum physics and chemistry
ER -
TY - JOUR
AB - We present a novel algorithm that allows us to gain detailed insight into the effects of sparsity in linear and nonlinear optimization, which is of great importance in many scientific areas such as image and signal processing, medical imaging, compressed sensing, and machine learning (e.g., for the training of neural networks). Sparsity is an important feature to ensure robustness against noisy data, but also to find models that are interpretable and easy to analyze due to the small number of relevant terms. It is common practice to enforce sparsity by adding the ℓ1-norm as a weighted penalty term. In order to gain a better understanding and to allow for an informed model selection, we directly solve the corresponding multiobjective optimization problem (MOP) that arises when we minimize the main objective and the ℓ1-norm simultaneously. As this MOP is in general non-convex for nonlinear objectives, the weighting method will fail to provide all optimal compromises. To avoid this issue, we present a continuation method which is specifically tailored to MOPs with two objective functions one of which is the ℓ1-norm. Our method can be seen as a generalization of well-known homotopy methods for linear regression problems to the nonlinear case. Several numerical examples - including neural network training - demonstrate our theoretical findings and the additional insight that can be gained by this multiobjective approach.
AU - Bieker, Katharina
AU - Gebken, Bennet
AU - Peitz, Sebastian
ID - 20731
JF - IEEE Transactions on Pattern Analysis and Machine Intelligence
TI - On the Treatment of Optimization Problems with L1 Penalty Terms via Multiobjective Continuation
ER -
TY - GEN
AB - Hamiltonian systems are differential equations which describe systems in classical mechanics, plasma physics, and sampling problems. They exhibit many structural properties, such as a lack of attractors and the presence of conservation laws. To predict Hamiltonian dynamics based on discrete trajectory observations, incorporation of prior knowledge about Hamiltonian structure greatly improves predictions. This is typically done by learning the system's Hamiltonian and then integrating the Hamiltonian vector field with a symplectic integrator. For this, however, Hamiltonian data needs to be approximated based on the trajectory observations. Moreover, the numerical integrator introduces an additional discretisation error. In this paper, we show that an inverse modified Hamiltonian structure adapted to the geometric integrator can be learned directly from observations. A separate approximation step for the Hamiltonian data avoided. The inverse modified data compensates for the discretisation error such that the discretisation error is eliminated.
AU - Offen, Christian
AU - Ober-Blöbaum, Sina
ID - 23382
TI - Symplectic integration of learned Hamiltonian systems
ER -