BZPEERPapers, people, evidence and research paths
Menu

Discover

SearchFind papers, people, topics, institutions and opportunities from one place.GraphSee how papers, authors, topics and evidence connect.

Research tools

HomeContinue from the papers, people and topics that matter now.CollectionsTurn saved papers into reading lists, evidence maps and shareable context.ResearchReconstruct the path that led you from search to paper to question.CollaborationMove from a paper, topic or expert request into a focused thread.PublishCreate full-text research articles with metadata, formulas and versions.

Network

ExpertsFind specialists whose work and reviews explain why they can help.CommunitiesJoin research groups organized around papers, topics and shared evidence.

Opportunities

ListingsBrowse roles, calls and collaborations connected to real research context.

Account

Log inReturn to your saved papers, graph views and active threads.Create accountStart saving papers, building a research trail and joining teams.
Log inCreate account

Source author record

Sebastian Schmitt

Sebastian Schmitt appears in the imported research catalog. Authorship, coauthor and topic links are available while profile ownership is still unclaimed.

ResearcherUnclaimed source record
cond-mat.str-elNeural and Evolutionary Computingcond-mat.mes-hallMachine LearningArtificial IntelligenceComputational Engineering, Finance, and ScienceEmerging Technologies

Catalog footprint

What is connected

16works
7topics
4close collaborators

Actions

Connect this record

Log in to claim
Open graphBrowse works

Research graph

See the researcher in context

Open full explorer

Inspect adjacent papers, topics, institutions and collaborators without losing the researcher page.

Building this map preview

BZPEER is loading the nearby papers, people, topics and institutions for this page.

Published work

16 published item(s)

preprint2022arXiv

A Scalable Approach to Modeling on Accelerated Neuromorphic Hardware

Neuromorphic systems open up opportunities to enlarge the explorative space for computational research. However, it is often challenging to unite efficiency and usability. This work presents the software aspects of this endeavor for the BrainScaleS-2 system, a hybrid accelerated neuromorphic hardware architecture based on physical modeling. We introduce key aspects of the BrainScaleS-2 Operating System: experiment workflow, API layering, software design, and platform operation. We present use cases to discuss and derive requirements for the software and showcase the implementation. The focus lies on novel system and software features such as multi-compartmental neurons, fast re-configuration for hardware-in-the-loop training, applications for the embedded processors, the non-spiking operation mode, interactive platform access, and sustainable hardware/software co-development. Finally, we discuss further developments in terms of hardware scale-up, system usability and efficiency.

0 citations0 reviewsNo code linkedOpen access
preprint2022arXiv

Alleviating Search Bias in Bayesian Evolutionary Optimization with Many Heterogeneous Objectives

Multi-objective optimization problems whose objectives have different evaluation costs are commonly seen in the real world. Such problems are now known as multi-objective optimization problems with heterogeneous objectives (HE-MOPs). So far, however, only a few studies have been reported to address HE-MOPs, and most of them focus on bi-objective problems with one fast objective and one slow objective. In this work, we aim to deal with HE-MOPs having more than two black-box and heterogeneous objectives. To this end, we develop a multi-objective Bayesian evolutionary optimization approach to HE-MOPs by exploiting the different data sets on the cheap and expensive objectives in HE-MOPs to alleviate the search bias caused by the heterogeneous evaluation costs for evaluating different objectives. To make the best use of two different training data sets, one with solutions evaluated on all objectives and the other with those only evaluated on the fast objectives, two separate Gaussian process models are constructed. In addition, a new acquisition function that mitigates search bias towards the fast objectives is suggested, thereby achieving a balance between convergence and diversity. We demonstrate the effectiveness of the proposed algorithm by testing it on widely used multi-/many-objective benchmark problems whose objectives are assumed to be heterogeneously expensive.

0 citations0 reviewsNo code linkedOpen access
preprint2022arXiv

Concept Identification for Complex Engineering Datasets

Finding meaningful concepts in engineering application datasets which allow for a sensible grouping of designs is very helpful in many contexts. It allows for determining different groups of designs with similar properties and provides useful knowledge in the engineering decision making process. Also, it opens the route for further refinements of specific design candidates which exhibit certain characteristic features. In this work, an approach to define meaningful and consistent concepts in an existing engineering dataset is presented. The designs in the dataset are characterized by a multitude of features such as design parameters, geometrical properties or performance values of the design for various boundary conditions. In the proposed approach the complete feature set is partitioned into several subsets called description spaces. The definition of the concepts respects this partitioning which leads to several desired properties of the identified concepts. This cannot be achieved with state-of-the-art clustering or concept identification approaches. A novel concept quality measure is proposed, which provides an objective value for a given definition of concepts in a dataset. The usefulness of the measure is demonstrated by considering a realistic engineering dataset consisting of about 2500 airfoil profiles, for which the performance values (lift and drag) for three different operating conditions were obtained by a computational fluid dynamics simulation. A numerical optimization procedure is employed, which maximizes the concept quality measure and finds meaningful concepts for different setups of the description spaces, while also incorporating user preference. It is demonstrated how these concepts can be used to select archetypal representatives of the dataset which exhibit characteristic features of each concept.

0 citations0 reviewsNo code linkedOpen access
preprint2022arXiv

Finding Efficient Trade-offs in Multi-Fidelity Response Surface Modeling

In the context of optimization approaches to engineering applications, time-consuming simulations are often utilized which can be configured to deliver solutions for various levels of accuracy, commonly referred to as different fidelity levels. It is common practice to train hierarchical surrogate models on the objective functions in order to speed-up the optimization process. These operate under the assumption that there is a correlation between the high- and low-fidelity versions of the problem that can be exploited to cheaply gain information. In the practical scenario where the computational budget has to be allocated between multiple fidelities, limited guidelines are available to help make that division. In this paper we evaluate a range of different choices for a two-fidelity setup that provide helpful intuitions about the trade-off between evaluating in high- or low-fidelity. We present a heuristic method based on subsampling from an initial Design of Experiments (DoE) to find a suitable division of the computational budget between the fidelity levels. This enables the setup of multi-fidelity optimizations which utilize the available computational budget efficiently, independent of the multi-fidelity model used.

0 citations0 reviewsNo code linkedOpen access
preprint2022arXiv

Stream-based Active Learning with Verification Latency in Non-stationary Environments

Data stream classification is an important problem in the field of machine learning. Due to the non-stationary nature of the data where the underlying distribution changes over time (concept drift), the model needs to continuously adapt to new data statistics. Stream-based Active Learning (AL) approaches address this problem by interactively querying a human expert to provide new data labels for the most recent samples, within a limited budget. Existing AL strategies assume that labels are immediately available, while in a real-world scenario the expert requires time to provide a queried label (verification latency), and by the time the requested labels arrive they may not be relevant anymore. In this article, we investigate the influence of finite, time-variable, and unknown verification delay, in the presence of concept drift on AL approaches. We propose PRopagate (PR), a latency independent utility estimator which also predicts the requested, but not yet known, labels. Furthermore, we propose a drift-dependent dynamic budget strategy, which uses a variable distribution of the labelling budget over time, after a detected drift. Thorough experimental evaluation, with both synthetic and real-world non-stationary datasets, and different settings of verification latency and budget are conducted and analyzed. We empirically show that the proposed method consistently outperforms the state-of-the-art. Additionally, we demonstrate that with variable budget allocation in time, it is possible to boost the performance of AL strategies, without increasing the overall labeling budget.

0 citations0 reviewsNo code linkedOpen access
preprint2022arXiv

The Operating System of the Neuromorphic BrainScaleS-1 System

BrainScaleS-1 is a wafer-scale mixed-signal accelerated neuromorphic system targeted for research in the fields of computational neuroscience and beyond-von-Neumann computing. The BrainScaleS Operating System (BrainScaleS OS) is a software stack giving users the possibility to emulate networks described in the high-level network description language PyNN with minimal knowledge of the system. At the same time, expert usage is facilitated by allowing to hook into the system at any depth of the stack. We present operation and development methodologies implemented for the BrainScaleS-1 neuromorphic architecture and walk through the individual components of BrainScaleS OS constituting the software stack for BrainScaleS-1 platform operation.

0 citations0 reviewsNo code linkedOpen access
preprint2020arXiv

Accelerated physical emulation of Bayesian inference in spiking neural networks

The massively parallel nature of biological information processing plays an important role for its superiority to human-engineered computing devices. In particular, it may hold the key to overcoming the von Neumann bottleneck that limits contemporary computer architectures. Physical-model neuromorphic devices seek to replicate not only this inherent parallelism, but also aspects of its microscopic dynamics in analog circuits emulating neurons and synapses. However, these machines require network models that are not only adept at solving particular tasks, but that can also cope with the inherent imperfections of analog substrates. We present a spiking network model that performs Bayesian inference through sampling on the BrainScaleS neuromorphic platform, where we use it for generative and discriminative computations on visual data. By illustrating its functionality on this platform, we implicitly demonstrate its robustness to various substrate-specific distortive effects, as well as its accelerated capability for computation. These results showcase the advantages of brain-inspired physical computation and provide important building blocks for large-scale neuromorphic applications.

0 citations0 reviewsNo code linkedOpen access
preprint2012arXiv

Extension of dynamical mean-field theory by inclusion of nonlocal two-site correlations with variable distance

We present a novel approximation scheme for the treatment of strongly correlated electrons in arbitrary crystal lattices. The approach extends the well-known dynamical mean field theory to include nonlocal two-site correlations of arbitrary spatial extent. We extract the nonlocal correlation functions from two-impurity Anderson models where the impurity-impurity distance defines the spatial extent of the correlations included. Translational invariance is fully respected by our approach since correlation functions of any two-impurity cluster are periodically embedded to $k$-space via a Fourier transform. As a first application, we study the two-dimensional Hubbard model on a simple-cubic lattice. We demonstrate how pseudogap formation in the many-body resonance at the Fermi level results from the inclusion of nonlocal correlations.

0 citations0 reviewsNo code linkedOpen access
preprint2012arXiv

Itinerant and local-moment magnetism in strongly correlated electron systems

Detailed analysis of the magnetic properties of the Hubbard model within dynamical mean-field theory (DMFT) is presented. Using a RPA-like decoupling of two-particle propagators we derive a universal form for susceptibilities, which captures essential aspects of localized and itinerant pictures. This expression is shown to be quantitatively valid whenever long-range coherence of particle-hole excitations can be neglected, as is the case in large parts of the phase diagram where antiferromag- netism is dominant. The applicability of an interpretation in terms of the two archetypical pictures of magnetism is investigated for the Hubbard model on a body-centered cubic lattice with additional next-nearest neighbor hopping t'. For large values of the Coulomb interaction, local-moment mag- netism is found to be dominant, while for weakly interacting band electrons itinerant quasiparticle magnetism prevails. In the intermediate regime and for finite t' an re-entrant behavior is discovered, where antiferromagnetism only exists in a finite temperature interval.

0 citations0 reviewsNo code linkedOpen access
preprint2012arXiv

Spectral properties of the two-impurity Anderson model with varying distance and various interactions

We present a novel approximation for two interacting magnetic impurities immersed into a noninteracting metallic host. The scheme is based on direct perturbation theory in the hybridization between the impurity and band electrons. This two-impurity enhanced noncrossing approximation (TIE) fully incorporates the indirect interaction between the impurities which are mediated by the conduction electrons as well as additional arbitrary direct interaction matrix elements. We qualify the approximation by investigating the uncoupled case and conclude the TIE to be equally accurate as its single impurity counterpart. The physical properties of the two-impurity Anderson model are first investigated in some limiting cases. In a regime, where each of the uncoupled two impurities would exhibit a pronounced Kondo effect, we ignore the indirect coupling via the conduction band and only incorporate direct interactions. For a ferromagnetic direct exchange coupling, the system displays a behavior similar to a spin-one Kondo effect, while an antiferromagnetic coupling competes with the Kondo effect and produces a pseudogap in the many-body Kondo resonance of the single-particle spectral function. Interestingly, a direct one-particle hopping also produces a pseudogap but additionally pronounced side-peaks emerge. This gap is characteristically different from the case with antiferromagnetic coupling, since it emerges as a consequence of distinct Kondo effects for the bonding and anti-bonding orbital, i.e.\ it reflects a splitting of even and odd parity states.For the general case of only indirect coupling the results show signatures of all the previously discussed limiting cases as function of the impurity-impurity distance. We are led to the conclusion that the well known Doniach-scenario captures essential aspects this model, but the details especially at small distances are more complicated.

0 citations0 reviewsNo code linkedOpen access
preprint2011arXiv

Comment on "Scaling feature of magnetic field induced Kondo-peak splittings"

In a recent work Zhang and coworkers (PRB 82, 075111 (2010)) studied the Zeeman splitting of the Kondo resonance for the single impurity Anderson model in a finite magnetic field $B$ with the numerical renormalization group (NRG) method. There, it was found that with increasing magnetic field $B$ the position of the Kondo resonance in the total spectral function \textit{does not} approach its position in the spin resolved spectral function. Additionally, the position of the Kondo maximum exceeded the Zeeman energy for $B/ T_K\gtrsim 5-10$, where $T_K$ is the low energy Kondo scale of the model ($g=2$, $μ_B=k_B=\hbar=1$). In this comment we argue that both these findings are produced by an improper choice of NRG parameter values. However, we reproduce the crossover in the splitting from Kondo-like behavior to a non-universal splitting larger than the Zeeman energy, but this crossover occurs at much larger fields of the order of the charge scale.

0 citations0 reviewsNo code linkedOpen access
preprint2011arXiv

Kinks in the electronic dispersion of the Hubbard model away from half filling

We study kinks in the electronic dispersion of a generic strongly correlated system by dynamic mean-field theory (DMFT). The focus is on doped systems away from particle-hole symmetry where valence fluctuations matter potentially. Three different algorithms are compared to asses their strengths and weaknesses, as well as to clearly distinguish physical features from algorithmic artifacts. Our findings extend a view previously established for half-filled systems where kinks reflect the coupling of the fermionic quasiparticles to emergent collective modes, which are identified here as spin fluctuations. Kinks are observed when strong spin fluctuations are present and, additionally, a separation of energy scales for spin and charge excitations exists. Both criteria are met by strongly correlated systems close to a Mott-insulator transition. The energies of the kinks and their doping dependence fit well to the kinks in the cuprates, which is surprising in view of the spatial correlations neglected by DMFT.

0 citations0 reviewsNo code linkedOpen access
preprint2011arXiv

Nonequilibrium Zeeman-splitting in quantum transport through nanoscale junctions

We calculate the nonequilibrium differential conductance $G(V)$ through a quantum dot as function of bias voltage $V$ and applied magnetic field $H$. We use a Keldysh conserving approximation for weakly correlated and the scattering states numerical renormalization group for the intermediate and strongly correlated regime out of equilibrium. In the weakly correlated regime, the Zeeman splitting observable in $G(V)$ strongly depends on the asymmetry of the coupling to the two leads, as well as on particle-hole asymmetry of the quantum dot. In contrast, in the strongly correlated regime, where Kondo-correlations dominate, the position $Δ_K$ of the Zeeman-split zero-bias anomaly is independent of such asymmetries and always found to be of the order of the Zeeman energy $Δ_0$. We find a crossover from the purely spin-fluctuation driven Kondo regime at small magnetic fields with $Δ_K<Δ_0$ to a regime at large fields where the contribution of charge fluctuations induces larger splittings with $Δ_K>Δ_0$ as it was observed in recent experiments.

0 citations0 reviewsNo code linkedOpen access
preprint2010arXiv

Comparison between scattering-states numerical renormalization group and the Kadanoff-Baym-Keldysh approach to quantum transport: Crossover from weak to strong correlations

The quantum transport through nanoscale junctions is governed by the charging energy $U$ of the device. We employ the recently developed scattering-states numerical renormalization group approach to open quantum systems to study nonequilibrium Green functions and current-voltage characteristics of such junctions for small and intermediate values of $U$. The reliability of the approach is established by the excellent agreement with diagrammatic Kadanoff-Baym-Keldysh results at small values of the $U$. We demonstrate the limits of the diagrammatic approaches at intermediate Coulomb repulsion. These approaches predict two different low-energy scale for magnetic and charge fluctuations in zero bias while the numerical renormalization group approach correctly yields only one single, universal scale. At large voltages and intermediate values of the Coulomb repulsion the self-consistent second Born as well as the GW approximation reproduce the SNRG spectral functions quite well for a symmetric junctions, while for the asymmetric model the voltage-dependent redistribution of spectral weight differs significantly. The second-order perturbation theory does not capture the correct single-particle dynamics at large bias and violates current conservation for asymmetric junctions.

0 citations0 reviewsNo code linkedOpen access
preprint2010arXiv

Dynamic susceptibilities of the single impurity Anderson model within an enhanced non-crossing approximation

The single impurity Anderson model (SIAM) is studied within an enhanced non-crossing approximation (ENCA). This method is extended to the calculation of susceptibilities and thoroughly tested, also in order to prepare applications as a building block for the calculation of susceptibilities and phase transitions in correlated lattice systems. A wide range of model parameters, such as impurity occupancy, temperature, local Coulomb repulsion and hybridization strength, are studied. Results for the spin and charge susceptibilities are presented. By comparing the static quantities to exact Bethe ansatz results, it is shown that the description of the magnetic excitations of the impurity within the ENCA is excellent, even in situations with large valence fluctuations or vanishing Coulomb repulsion. The description of the charge susceptibility is quite accurate in situations where the singly occupied ionic configuration is the unperturbed ground state; however, it seems to overestimate charge fluctuations in the asymmetric model at too low temperatures. The dynamic spin excitation spectra is dominated by the Kondo-screening of the impurity spin through the conduction band, i.e. the formation of the local Kondo-singlet. A finite local Coulomb interaction U leads to a drastic reduction of the charge response via processes involving the doubly occupied impurity state. In the asymmetric model, the charge susceptibility is enhanced for excitation energies smaller than the Kondo scale T_K due to the influence of valence fluctuations.

0 citations0 reviewsNo code linkedOpen access
preprint2010arXiv

Non-Fermi liquid signatures in the Hubbard Model due to van Hove singularities

When a van-Hove singularity is located in the vicinity of the Fermi level, the electronic scattering rate acquires a non-analytic contribution. This invalidates basic assumptions of Fermi liquid theory and within perturbative treatments leads to a non-Fermi liquid self-energy and transport properties.Such anomalies are shown to also occur in the strongly correlated metallic state. We consider the Hubbard model on a two-dimensional square lattice with nearest and next-nearest neighbor hopping within the single-site dynamical mean-field theory. At temperatures on the order of the low-energy scale $T_0$ an unusual maximum emerges in the imaginary part of the self-energy which is renormalized towards the Fermi level for finite doping. At zero temperature this double-well structure is suppressed, but an anomalous energy dependence of the self-energy remains. For the frustrated Hubbard model on the square lattice with next-nearest neighbor hopping, the presence of the van Hove singularity changes the asymptotic low temperature behavior of the resistivity from a Fermi liquid to non-Fermi liquid dependency as function of doping. The results of this work are discussed regarding their relevance for high-temperature cuprate superconductors.

0 citations0 reviewsNo code linkedOpen access