Showing 1–12 of 12 results for author: Frenzel, M

Landscape-Aware Automated Algorithm Configuration using Multi-output Mixed Regression and Classification

Authors: Fu Xing Long, Moritz Frenzel, Peter Krause, Markus Gitterle, Thomas Bäck, Niki van Stein

Abstract: In landscape-aware algorithm selection problem, the effectiveness of feature-based predictive models strongly depends on the representativeness of training data for practical applications. In this work, we investigate the potential of randomly generated functions (RGF) for the model training, which cover a much more diverse set of optimization problem classes compared to the widely-used black-box… ▽ More In landscape-aware algorithm selection problem, the effectiveness of feature-based predictive models strongly depends on the representativeness of training data for practical applications. In this work, we investigate the potential of randomly generated functions (RGF) for the model training, which cover a much more diverse set of optimization problem classes compared to the widely-used black-box optimization benchmarking (BBOB) suite. Correspondingly, we focus on automated algorithm configuration (AAC), that is, selecting the best suited algorithm and fine-tuning its hyperparameters based on the landscape features of problem instances. Precisely, we analyze the performance of dense neural network (NN) models in handling the multi-output mixed regression and classification tasks using different training data sets, such as RGF and many-affine BBOB (MA-BBOB) functions. Based on our results on the BBOB functions in 5d and 20d, near optimal configurations can be identified using the proposed approach, which can most of the time outperform the off-the-shelf default configuration considered by practitioners with limited knowledge about AAC. Furthermore, the predicted configurations are competitive against the single best solver in many cases. Overall, configurations with better performance can be best identified by using NN models trained on a combination of RGF and MA-BBOB functions. △ Less

Submitted 2 September, 2024; originally announced September 2024.

Electro-Optic Cavities for In-Situ Measurement of Cavity Fields

Authors: Michael S. Spencer, Joanna M. Urban, Maximilian Frenzel, Niclas S. Mueller, Olga Minakova, Martin Wolf, Alexander Paarmann, Sebastian F. Maehrlein

Abstract: Cavity electrodynamics offers a unique avenue for tailoring ground-state material properties, excited-state engineering, and versatile control of quantum matter. Merging these concepts with high-field physics in the terahertz (THz) spectral range opens the door to explore low-energy, field-driven cavity electrodynamics, emerging from fundamental resonances or order parameters. Despite this demand,… ▽ More Cavity electrodynamics offers a unique avenue for tailoring ground-state material properties, excited-state engineering, and versatile control of quantum matter. Merging these concepts with high-field physics in the terahertz (THz) spectral range opens the door to explore low-energy, field-driven cavity electrodynamics, emerging from fundamental resonances or order parameters. Despite this demand, leveraging the full potential of field-driven material control in cavities is hindered by the lack of direct access to the intra-cavity fields. Here, we demonstrate a new concept of active cavities, consisting of electro-optic Fabry-Perot resonators, which measure their intra-cavity electric fields on sub-cycle timescales. We thereby demonstrate quantitative retrieval of the cavity modes in amplitude and phase, over a broad THz frequency range. To enable simultaneous intra-cavity sampling alongside excited-state material control, we design a tunable multi-layer cavity, enabling deterministic design of hybrid cavities for polaritonic systems. Our theoretical models reveal the origin of the avoided crossings embedded in the intricate mode dispersion, and will enable fully-switchable polaritonic effects within arbitrary materials hosted by the hybrid cavity. Electro-optic cavities (EOCs) will therefore serve as integrated probes of light-matter interactions across all coupling regimes, laying the foundation for field-resolved intra-cavity quantum electrodynamics. △ Less

Submitted 20 June, 2024; originally announced June 2024.

Comments: main: 7 pages, 4 figurs; total: 34 pages, 20 figures

Quartz as an Accurate High-Field Low-Cost THz Helicity Detector

Authors: Maximilian Frenzel, Joanna M. Urban, Leona Nest, Tobias Kampfrath, Michael S. Spencer, Sebastian F. Maehrlein

Abstract: The advent of high-field THz sources has opened the field of nonlinear THz physics and unlocked access to fundamental low energy excitations for ultrafast material control. Recent advances towards controlling and employing chiral excitations, or generally angular momentum of light, not only rely on the measurement of undistorted intense THz fields, but also on the precise knowledge about sophistic… ▽ More The advent of high-field THz sources has opened the field of nonlinear THz physics and unlocked access to fundamental low energy excitations for ultrafast material control. Recent advances towards controlling and employing chiral excitations, or generally angular momentum of light, not only rely on the measurement of undistorted intense THz fields, but also on the precise knowledge about sophisticated THz helicity states. A recently reported and promising detector material is $α$-quartz. However, its electrooptic response function and contributing nonlinear effects have remained elusive. Here, we establish z-cut $α$-quartz as a precise electrooptic THz detector for full amplitude, phase and polarization measurement of intense THz fields, all at a fraction of costs of conventional THz detectors. We experimentally determine its complex detector response function, which is in good agreement with our model based on predominantly known literature values. It also explains previously observed thickness-dependent waveforms. These insights allow us to develop a swift and reliable protocol to precisely measure arbitrary THz polarization and helicity states. This two-dimensional electrooptic sampling (2D-EOS) in $α$-quartz fosters rapid and cost-efficient THz time-domain ellipsometry, and enables the characterization of polarization-tailored fields for driving chiral or other helicity-sensitive quasiparticles and topologies. △ Less

Submitted 15 September, 2023; originally announced September 2023.

Comments: 26 pages, 9 figures, 1 table

Synthesising Rare Cataract Surgery Samples with Guided Diffusion Models

Authors: Yannik Frisch, Moritz Fuchs, Antoine Sanner, Felix Anton Ucar, Marius Frenzel, Joana Wasielica-Poslednik, Adrian Gericke, Felix Mathias Wagner, Thomas Dratsch, Anirban Mukhopadhyay

Abstract: Cataract surgery is a frequently performed procedure that demands automation and advanced assistance systems. However, gathering and annotating data for training such systems is resource intensive. The publicly available data also comprises severe imbalances inherent to the surgical process. Motivated by this, we analyse cataract surgery video data for the worst-performing phases of a pre-trained… ▽ More Cataract surgery is a frequently performed procedure that demands automation and advanced assistance systems. However, gathering and annotating data for training such systems is resource intensive. The publicly available data also comprises severe imbalances inherent to the surgical process. Motivated by this, we analyse cataract surgery video data for the worst-performing phases of a pre-trained downstream tool classifier. The analysis demonstrates that imbalances deteriorate the classifier's performance on underrepresented cases. To address this challenge, we utilise a conditional generative model based on Denoising Diffusion Implicit Models (DDIM) and Classifier-Free Guidance (CFG). Our model can synthesise diverse, high-quality examples based on complex multi-class multi-label conditions, such as surgical phases and combinations of surgical tools. We affirm that the synthesised samples display tools that the classifier recognises. These samples are hard to differentiate from real images, even for clinical experts with more than five years of experience. Further, our synthetically extended data can improve the data sparsity problem for the downstream task of tool classification. The evaluations demonstrate that the model can generate valuable unseen examples, allowing the tool classifier to improve by up to 10% for rare cases. Overall, our approach can facilitate the development of automated assistance systems for cataract surgery by providing a reliable source of realistic synthetic data, which we make available for everyone. △ Less

Submitted 3 August, 2023; originally announced August 2023.

DoE2Vec: Deep-learning Based Features for Exploratory Landscape Analysis

Authors: Bas van Stein, Fu Xing Long, Moritz Frenzel, Peter Krause, Markus Gitterle, Thomas Bäck

Abstract: We propose DoE2Vec, a variational autoencoder (VAE)-based methodology to learn optimization landscape characteristics for downstream meta-learning tasks, e.g., automated selection of optimization algorithms. Principally, using large training data sets generated with a random function generator, DoE2Vec self-learns an informative latent representation for any design of experiments (DoE). Unlike the… ▽ More We propose DoE2Vec, a variational autoencoder (VAE)-based methodology to learn optimization landscape characteristics for downstream meta-learning tasks, e.g., automated selection of optimization algorithms. Principally, using large training data sets generated with a random function generator, DoE2Vec self-learns an informative latent representation for any design of experiments (DoE). Unlike the classical exploratory landscape analysis (ELA) method, our approach does not require any feature engineering and is easily applicable for high dimensional search spaces. For validation, we inspect the quality of latent reconstructions and analyze the latent representations using different experiments. The latent representations not only show promising potentials in identifying similar (cheap-to-evaluate) surrogate functions, but also can significantly boost performances when being used complementary to the classical ELA features in classification tasks. △ Less

Submitted 31 March, 2023; originally announced April 2023.

Nonlinear THz Control of the Lead Halide Perovskite Lattice

Authors: Maximilian Frenzel, Marie Cherasse, Joanna M. Urban, Feifan Wang, Bo Xiang, Leona Nest, Lucas Huber, Luca Perfetti, Martin Wolf, Tobias Kampfrath, Xiaoyang Zhu, Sebatian F. Maehrlein

Abstract: Lead halide perovskites (LHPs) have emerged as an excellent class of semiconductors for next-generation solar cells and optoelectronic devices. Tailoring physical properties by fine-tuning the lattice structures has been explored in these materials by chemical composition or morphology. Nevertheless, its dynamic counterpart, phonon-driven ultrafast material control, as contemporarily harnessed for… ▽ More Lead halide perovskites (LHPs) have emerged as an excellent class of semiconductors for next-generation solar cells and optoelectronic devices. Tailoring physical properties by fine-tuning the lattice structures has been explored in these materials by chemical composition or morphology. Nevertheless, its dynamic counterpart, phonon-driven ultrafast material control, as contemporarily harnessed for oxide perovskites, has not been established yet. Here we employ intense THz electric fields to obtain direct lattice control via nonlinear excitation of coherent octahedral twist modes in hybrid CH3NH3PbBr3 and all-inorganic CsPbBr3 perovskites. These Raman-active phonons at 0.9 - 1.3 THz are found to govern the ultrafast THz-induced Kerr effect in the low-temperature orthorhombic phase and thus dominate the phonon-modulated polarizability with potential implications for dynamic charge carrier screening beyond the Froehlich polaron. Our work opens the door to selective control of LHP's vibrational degrees of freedom governing phase transitions and dynamic disorder. △ Less

Submitted 9 January, 2023; originally announced January 2023.

Comments: 11 pages, 5 figures

Report number: eadg3856

Journal ref: Science Advances , Vol. 9, No. 21, p. eadg3856 (2023)

Latent Space Cartography: Generalised Metric-Inspired Measures and Measure-Based Transformations for Generative Models

Authors: Max F. Frenzel, Bogdan Teleaga, Asahi Ushio

Abstract: Deep generative models are universal tools for learning data distributions on high dimensional data spaces via a mapping to lower dimensional latent spaces. We provide a study of latent space geometries and extend and build upon previous results on Riemannian metrics. We show how a class of heuristic measures gives more flexibility in finding meaningful, problem-specific distances, and how it can… ▽ More Deep generative models are universal tools for learning data distributions on high dimensional data spaces via a mapping to lower dimensional latent spaces. We provide a study of latent space geometries and extend and build upon previous results on Riemannian metrics. We show how a class of heuristic measures gives more flexibility in finding meaningful, problem-specific distances, and how it can be applied to diverse generator types such as autoregressive generators commonly used in e.g. language and other sequence modeling. We further demonstrate how a diffusion-inspired transformation previously studied in cartography can be used to smooth out latent spaces, stretching them according to a chosen measure. In addition to providing more meaningful distances directly in latent space, this also provides a unique tool for novel kinds of data visualizations. We believe that the proposed methods can be a valuable tool for studying the structure of latent spaces and learned data distributions of generative models. △ Less

Submitted 6 February, 2019; originally announced February 2019.

Coarse-grained hidden entropy production in partially inaccessible quantum jump trajectories

Authors: Max F. Frenzel, Takahiro Sagawa

Abstract: We consider an open quantum system for which only a subset of all possible transitions are accessible, while the remaining ones are hidden from direct observation. Using a modification of the notion of quantum jump trajectories we introduce the coarse-grained hidden entropy, which quantifies the entropy production in the hidden subsystem conditioned on our observations of the visible part. The ent… ▽ More We consider an open quantum system for which only a subset of all possible transitions are accessible, while the remaining ones are hidden from direct observation. Using a modification of the notion of quantum jump trajectories we introduce the coarse-grained hidden entropy, which quantifies the entropy production in the hidden subsystem conditioned on our observations of the visible part. The entropy production consisting of the sum of visible and coarse-grained hidden entropy is shown to satisfy an integral fluctuations theorem. Depending on the nature of the continuously occurring measurement and feedback processes between the interacting subsystems, the visible entropy can take negative values in which case the hidden systems acts as a Maxwell's demon. Our results allow us to study quantum-mechanical effects in autonomous systems, such as the autonomous quantum Maxwell's demon we introduce as an illustrative example. △ Less

Submitted 27 September, 2016; originally announced September 2016.

Quasi-autonomous quantum thermal machines and quantum to classical energy flow

Authors: Max F. Frenzel, David Jennings, Terry Rudolph

Abstract: There are both practical and foundational motivations to consider the thermodynamics of quantum systems at small scales. Here we address the issue of autonomous quantum thermal machines that are tailored to achieve some specific thermodynamic primitive, such as work extraction in the presence of a thermal environment, while having minimal or no control from the macroscopic regime. Beyond experimen… ▽ More There are both practical and foundational motivations to consider the thermodynamics of quantum systems at small scales. Here we address the issue of autonomous quantum thermal machines that are tailored to achieve some specific thermodynamic primitive, such as work extraction in the presence of a thermal environment, while having minimal or no control from the macroscopic regime. Beyond experimental implementations, this provides an arena in which to address certain foundational aspects such as the role of coherence in thermodynamics, the use of clock degrees of freedom and the simulation of local time-dependent Hamiltonians in a particular quantum subsystem. For small-scale systems additional issues arise. Firstly, it is not clear to what degree genuine ordered thermodynamic work has been extracted, and secondly non-trivial back-actions on the thermal machine must be accounted for. We find that both these aspects can be resolved through a judicious choice of quantum measurements that magnify thermodynamic properties up the ladder of length-scales, while simultaneously stabilizing the quantum thermal machine. Within this framework we show that thermodynamic reversibility is obtained in a particular Zeno limit, and finally illustrate these concepts with a concrete example involving spin-systems. △ Less

Submitted 15 November, 2015; v1 submitted 11 August, 2015; originally announced August 2015.

Comments: v2: Extended discussions, added references, fixed typos; v3: fixed references

Journal ref: New J. Phys. 18 023037 (2016)

Reexamination of Pure Qubit Work Extraction

Authors: Max F. Frenzel, David Jennings, Terry Rudolph

Abstract: Many work extraction or information erasure processes in the literature involve the raising and lowering of energy levels via external fields. But even if the actual system is treated quantum mechanically, the field is assumed to be classical and of infinite strength, hence not developing any correlations with the system or experiencing back-actions. We extend these considerations to a fully quant… ▽ More Many work extraction or information erasure processes in the literature involve the raising and lowering of energy levels via external fields. But even if the actual system is treated quantum mechanically, the field is assumed to be classical and of infinite strength, hence not developing any correlations with the system or experiencing back-actions. We extend these considerations to a fully quantum mechanical treatment, by studying a spin-1/2 particle coupled to a finite-sized directional quantum reference frame, a spin-l system, which models an external field. With this concrete model together with a bosonic thermal bath, we analyse the back-action a finite-size field suffers during a quantum-mechanical work extraction process, the effect this has on the extractable work, and highlight a range of assumptions commonly made when considering such processes. The well-known semi-classical treatment of work extraction from a pure qubit predicts a maximum extractable work W = kT log 2 for a quasi-static process, which holds as a strict upper bound in the fully quantum mechanical case, and is only attained in the classical limit. We also address the problem of emergent local time-dependence in a joint system with globally fixed Hamiltonian. △ Less

Submitted 19 November, 2014; v1 submitted 16 June, 2014; originally announced June 2014.

Comments: 16 pages, 11 figures; v2: updated figures and extended discussion; v3: changed title and minor updates, published in PRE

Journal ref: Phys. Rev. E 90, 052136 (2014)

Atoms and Photons - Their Interaction Dynamics

Authors: Max F. Frenzel

Abstract: The Jaynes-Cummings model (JCM) describes the interaction of a two-level atom with a single quantised field mode in an optical cavity. In this BSc project we derive the basic Rabi model Hamiltonian and show how it leads to the JCM Hamiltonian. The results are then used to analyse the general time evolution of a bipartite atom-field system and its properties, such as entanglement. We extend a schem… ▽ More The Jaynes-Cummings model (JCM) describes the interaction of a two-level atom with a single quantised field mode in an optical cavity. In this BSc project we derive the basic Rabi model Hamiltonian and show how it leads to the JCM Hamiltonian. The results are then used to analyse the general time evolution of a bipartite atom-field system and its properties, such as entanglement. We extend a scheme developed by Browne and Plenio [D. Browne and M. B. Plenio, Physical Review A, vol. 67, no. 1, Jan. 2003.] to entangle two cavities that interact with an atom. Our scheme allows to produce states that are arbitrarily close to a Bell state with a success probability greater than 0.75, whereas the success rate of the original method tends to 0 for perfect fidelity. Further, we extend both schemes by allowing the atom or the cavity to be in a thermal state and find that the fidelity of the produced states is reduced with increasing temperature. Finally, we apply our scheme to quantum teleportation, where we find that the fidelity of the teleported state compared to the original state also decreases as the temperature is increased. △ Less

Submitted 18 December, 2012; originally announced December 2012.

Matrix Product State Representation without explicit local Hilbert Space Truncation with Applications to the Sub-Ohmic Spin-Boson Model

Authors: Max F. Frenzel, Martin B. Plenio

Abstract: We present an alternative to the conventional matrix product state representation, which allows us to avoid the explicit local Hilbert space truncation many numerical methods employ. Utilising chain mappings corresponding to linear and logarithmic discretizations of the spin-boson model onto a semi-infinite chain, we apply the new method to the sub-ohmic SBM. We are able to reproduce many well-est… ▽ More We present an alternative to the conventional matrix product state representation, which allows us to avoid the explicit local Hilbert space truncation many numerical methods employ. Utilising chain mappings corresponding to linear and logarithmic discretizations of the spin-boson model onto a semi-infinite chain, we apply the new method to the sub-ohmic SBM. We are able to reproduce many well-established features of the quantum phase transition, such as the critical exponent 1/2 predicted by mean-field theory. Via extrapolation of finite-chain results, we are able to determine the infinite-chain critical couplings at which the transition occurs and, in general, study the behaviour of the system well into the localised phase. △ Less

Submitted 5 July, 2013; v1 submitted 14 November, 2012; originally announced November 2012.

Comments: 8 pages, 8 figures

Journal ref: New J. Phys. 15 073046 (2013)