-
Numerical methods for closed-loop systems with non-autonomous data
Authors:
B. Baran,
P. Benner,
J. Saak,
T. Stillfjord
Abstract:
By computing a feedback control via the linear quadratic regulator (LQR) approach and simulating a non-linear non-autonomous closed-loop system using this feedback, we combine two numerically challenging tasks. For the first task, the computation of the feedback control, we use the non-autonomous generalized differential Riccati equation (DRE), whose solution determines the time-varying feedback g…
▽ More
By computing a feedback control via the linear quadratic regulator (LQR) approach and simulating a non-linear non-autonomous closed-loop system using this feedback, we combine two numerically challenging tasks. For the first task, the computation of the feedback control, we use the non-autonomous generalized differential Riccati equation (DRE), whose solution determines the time-varying feedback gain matrix. Regarding the second task, we want to be able to simulate non-linear closed-loop systems for which it is known that the regulator is only valid for sufficiently small perturbations. Thus, one easily runs into numerical issues in the integrators when the closed-loop control varies greatly. For these systems, e.g., the A-stable implicit Euler methods fails.\newline On the one hand, we implement non-autonomous versions of splitting schemes and BDF methods for the solution of our non-autonomous DREs. These are well-established DRE solvers in the autonomous case. On the other hand, to tackle the numerical issues in the simulation of the non-linear closed-loop system, we apply a fractional-step-theta scheme with time-adaptivity tuned specifically to this kind of challenge. That is, we additionally base the time-adaptivity on the activity of the control. We compare this approach to the more classical error-based time-adaptivity.\newline We describe techniques to make these two tasks computable in a reasonable amount of time and are able to simulate closed-loop systems with strongly varying controls, while avoiding numerical issues. Our time-adaptivity approach requires fewer time steps than the error-based alternative and is more reliable.
△ Less
Submitted 21 February, 2024;
originally announced February 2024.
-
Meta-tuning Loss Functions and Data Augmentation for Few-shot Object Detection
Authors:
Berkan Demirel,
Orhun Buğra Baran,
Ramazan Gokberk Cinbis
Abstract:
Few-shot object detection, the problem of modelling novel object detection categories with few training instances, is an emerging topic in the area of few-shot learning and object detection. Contemporary techniques can be divided into two groups: fine-tuning based and meta-learning based approaches. While meta-learning approaches aim to learn dedicated meta-models for mapping samples to novel clas…
▽ More
Few-shot object detection, the problem of modelling novel object detection categories with few training instances, is an emerging topic in the area of few-shot learning and object detection. Contemporary techniques can be divided into two groups: fine-tuning based and meta-learning based approaches. While meta-learning approaches aim to learn dedicated meta-models for mapping samples to novel class models, fine-tuning approaches tackle few-shot detection in a simpler manner, by adapting the detection model to novel classes through gradient based optimization. Despite their simplicity, fine-tuning based approaches typically yield competitive detection results. Based on this observation, we focus on the role of loss functions and augmentations as the force driving the fine-tuning process, and propose to tune their dynamics through meta-learning principles. The proposed training scheme, therefore, allows learning inductive biases that can boost few-shot detection, while keeping the advantages of fine-tuning based approaches. In addition, the proposed approach yields interpretable loss functions, as opposed to highly parametric and complex few-shot meta-models. The experimental results highlight the merits of the proposed scheme, with significant improvements over the strong fine-tuning based few-shot detection baselines on benchmark Pascal VOC and MS-COCO datasets, in terms of both standard and generalized few-shot performance metrics.
△ Less
Submitted 24 April, 2023;
originally announced April 2023.
-
Riccati-feedback Control of a Two-dimensional Two-phase Stefan Problem
Authors:
Björn Baran,
Peter Benner,
Jens Saak
Abstract:
We discuss the feedback control problem for a two-dimensional two-phase Stefan problem. In our approach, we use a sharp interface representation in combination with mesh-movement to track the interface position. To attain a feedback control, we apply the linear-quadratic regulator approach to a suitable linearization of the problem. We address details regarding the discretization and the interface…
▽ More
We discuss the feedback control problem for a two-dimensional two-phase Stefan problem. In our approach, we use a sharp interface representation in combination with mesh-movement to track the interface position. To attain a feedback control, we apply the linear-quadratic regulator approach to a suitable linearization of the problem. We address details regarding the discretization and the interface representation therein. Further, we document the matrix assembly to generate a non-autonomous generalized differential Riccati equation. To numerically solve the Riccati equation, we use low-rank factored and matrix-valued versions of the non-autonomous backward differentiation formulas, which incorporate implicit index reduction techniques. For the numerical simulation of the feedback controlled Stefan problem, we use a time-adaptive fractional-step-theta scheme.
We provide the implementations for the developed methods and test these in several numerical experiments. With these experiments we show that our feedback control approach is applicable to the Stefan control problem and makes this large-scale problem computable. Also, we discuss the influence of several controller design parameters, such as the choice of inputs and outputs.
△ Less
Submitted 21 December, 2022; v1 submitted 12 September, 2022;
originally announced September 2022.
-
Semantics-driven Attentive Few-shot Learning over Clean and Noisy Samples
Authors:
Orhun Buğra Baran,
Ramazan Gökberk Cinbiş
Abstract:
Over the last couple of years few-shot learning (FSL) has attracted great attention towards minimizing the dependency on labeled training examples. An inherent difficulty in FSL is the handling of ambiguities resulting from having too few training samples per class. To tackle this fundamental challenge in FSL, we aim to train meta-learner models that can leverage prior semantic knowledge about nov…
▽ More
Over the last couple of years few-shot learning (FSL) has attracted great attention towards minimizing the dependency on labeled training examples. An inherent difficulty in FSL is the handling of ambiguities resulting from having too few training samples per class. To tackle this fundamental challenge in FSL, we aim to train meta-learner models that can leverage prior semantic knowledge about novel classes to guide the classifier synthesis process. In particular, we propose semantically-conditioned feature attention and sample attention mechanisms that estimate the importance of representation dimensions and training instances. We also study the problem of sample noise in FSL, towards the utilization of meta-learners in more realistic and imperfect settings. Our experimental results demonstrate the effectiveness of the proposed semantic FSL model with and without sample noise.
△ Less
Submitted 3 February, 2023; v1 submitted 9 January, 2022;
originally announced January 2022.
-
MaskSplit: Self-supervised Meta-learning for Few-shot Semantic Segmentation
Authors:
Mustafa Sercan Amac,
Ahmet Sencan,
Orhun Bugra Baran,
Nazli Ikizler-Cinbis,
Ramazan Gokberk Cinbis
Abstract:
Just like other few-shot learning problems, few-shot segmentation aims to minimize the need for manual annotation, which is particularly costly in segmentation tasks. Even though the few-shot setting reduces this cost for novel test classes, there is still a need to annotate the training data. To alleviate this need, we propose a self-supervised training approach for learning few-shot segmentation…
▽ More
Just like other few-shot learning problems, few-shot segmentation aims to minimize the need for manual annotation, which is particularly costly in segmentation tasks. Even though the few-shot setting reduces this cost for novel test classes, there is still a need to annotate the training data. To alleviate this need, we propose a self-supervised training approach for learning few-shot segmentation models. We first use unsupervised saliency estimation to obtain pseudo-masks on images. We then train a simple prototype based model over different splits of pseudo masks and augmentations of images. Our extensive experiments show that the proposed approach achieves promising results, highlighting the potential of self-supervised training. To the best of our knowledge this is the first work that addresses unsupervised few-shot segmentation problem on natural images.
△ Less
Submitted 3 November, 2021; v1 submitted 23 October, 2021;
originally announced October 2021.
-
Subgap dynamics of double quantum dot coupled between superconducting and normal leads
Authors:
B. Baran,
R. Taranko,
T. Domanski
Abstract:
Dynamical processes induced by the external time-dependent fields can provide valuable insight into the characteristic energy scales of a given physical system. We investigate them here in a nanoscopic heterostructure, consisting of the double quantum dot coupled in series to the superconducting and the metallic reservoirs, analyzing its response to (i)~abrupt bias voltage applied across the junct…
▽ More
Dynamical processes induced by the external time-dependent fields can provide valuable insight into the characteristic energy scales of a given physical system. We investigate them here in a nanoscopic heterostructure, consisting of the double quantum dot coupled in series to the superconducting and the metallic reservoirs, analyzing its response to (i)~abrupt bias voltage applied across the junction, (ii) sudden change of the energy levels, and imposed by (iii)~their periodic driving. We explore subgap properties of this setup which are strictly related to the in-gap quasiparticles and discuss their signatures manifested in the time-dependent charge currents. The characteristic multi-mode oscillations, their beating patters and photon-assisted harmonics reveal a rich spectrum of dynamical features that might be important for designing the superconducting qubits.
△ Less
Submitted 27 May, 2021;
originally announced May 2021.
-
Transient effects in double quantum dot sandwiched laterally between superconducting and metallic leads
Authors:
R. Taranko,
K. Wrześniewski,
B. Baran,
I. Weymann,
T. Domański
Abstract:
We study the transient phenomena appearing in a subgap region of the double quantum dot coupled in series between the superconducting and normal metallic leads, focusing on the development of the superconducting proximity effect. For the uncorrelated nanostructure we derive explicit expressions of the time-dependent occupancies in both quantum dots, charge currents, and electron pairing induced on…
▽ More
We study the transient phenomena appearing in a subgap region of the double quantum dot coupled in series between the superconducting and normal metallic leads, focusing on the development of the superconducting proximity effect. For the uncorrelated nanostructure we derive explicit expressions of the time-dependent occupancies in both quantum dots, charge currents, and electron pairing induced on individual dots and between them. We show that the initial configurations substantially affect the dynamical processes, in which the in-gap bound states emerge upon coupling the double quantum dot to superconducting reservoir. In particular, the superconducting proximity effect would be temporarily blocked whenever the quantum dots are initially singly occupied. Such {\it triplet}/{\it Andreev blockade} has been recently reported experimentally for double quantum dots embedded in the Josephson [D. Bouman et al., Phys. Rev. B 102, 220505 (2020)] and Andreev [P. Zhang et al., arXiv:2102.03283 (2021)] junctions. We also address the role of correlation effects within the lowest-order decoupling scheme and by the time-dependent numerical renormalization group calculations. Competition of the repulsive Coulomb interactions with the superconducting proximity effect leads to renormalization of the in-gap quasiparticles, speeding up the quantum oscillations and narrowing a region of transient phenomena, whereas the dynamical Andreev blockade is well pronounced in the weak inter-dot coupling limit. We propose feasible methods for detecting the characteristic time-scales that could be observable by the Andreev spectroscopy.
△ Less
Submitted 30 April, 2021;
originally announced April 2021.
-
Quench dynamics of correlated quantum dot proximitized to superconducting lead
Authors:
K. Wrześniewski,
B. Baran,
R. Taranko,
T. Domański,
I. Weymann
Abstract:
Quantum system abruptly driven from its stationary phase can reveal nontrivial dynamics upon approaching a new final state. We investigate here such dynamics for a correlated quantum dot sandwiched between the metallic and superconducting leads, considering two types of quenches feasible experimentally. In particular, we examine an interplay between the proximity induced electron pairing with corr…
▽ More
Quantum system abruptly driven from its stationary phase can reveal nontrivial dynamics upon approaching a new final state. We investigate here such dynamics for a correlated quantum dot sandwiched between the metallic and superconducting leads, considering two types of quenches feasible experimentally. In particular, we examine an interplay between the proximity induced electron pairing with correlations caused by the on-dot Coulomb repulsion. We discuss the time-dependent charge occupancy, complex order parameter, transient currents, and analyze evolution of the subgap quasiparticles which could be empirically observed in the tunneling conductance.
△ Less
Submitted 21 July, 2020;
originally announced July 2020.
-
A Multi-Objective Approach for Multi-Cloud Infrastructure Brokering in Dynamic Markets
Authors:
Fabio Lopez-Pires,
Lino Chamorro,
Benjamin Baran
Abstract:
Cloud Service Brokers (CSBs) facilitate complex resource allocation decisions, efficiently mapping dynamic tenant demands onto dynamic provider offers, where several objectives should ideally be considered. This work proposes for the first time a pure multi-objective formulation of a broker-oriented Virtual Machine Placement (VMP) problem for dynamic environments, simultaneously optimizing the fol…
▽ More
Cloud Service Brokers (CSBs) facilitate complex resource allocation decisions, efficiently mapping dynamic tenant demands onto dynamic provider offers, where several objectives should ideally be considered. This work proposes for the first time a pure multi-objective formulation of a broker-oriented Virtual Machine Placement (VMP) problem for dynamic environments, simultaneously optimizing the following objective functions: (i) Total Infrastructure CPU (TICPU), (ii) Total Infrastructure Memory (TIMEM) and (iii) Total Infrastructure Price (TIP) while considering load balancing across providers. To solve the formulated multi-objective problem, a Multi-Objective Evolutionary Algorithm (MOEA) is proposed. Considering that each time a demand (or offer) change occurs, a set of non-dominated solutions is found by Pareto-based algorithms as the one proposed, different selection strategies were evaluated in order to automatically select a convenient solution. Additionally, the proposed algorithm, including the considered selection strategies, was compared against mono-objective state-of-the-art alternatives in different scenarios with real data from providers in actual markets. Experimental results demonstrate that a pure multi-objective optimization approach considering the preferred solution selection strategy (S3) outperformed other mono-objective evaluated alternatives.
△ Less
Submitted 8 January, 2020;
originally announced January 2020.
-
Quasiparticles of periodically driven quantum dot coupled between superconducting and normal leads
Authors:
B. Baran,
T. Domański
Abstract:
We investigate subgap quasiparticles of a single level quantum dot coupled to the superconducting and normal leads, whose energy level is periodically driven by external potential. Using the Floquet formalism we determine the quasienergies and analyze redistribution of their spectral weights between individual harmonics upon varying the frequency and amplitude of the driving potential. We also pro…
▽ More
We investigate subgap quasiparticles of a single level quantum dot coupled to the superconducting and normal leads, whose energy level is periodically driven by external potential. Using the Floquet formalism we determine the quasienergies and analyze redistribution of their spectral weights between individual harmonics upon varying the frequency and amplitude of the driving potential. We also propose feasible spectroscopic methods for probing the in-gap quasiparticles observable in the differential conductance of the charge current averaged over a period of oscillations.
△ Less
Submitted 25 March, 2019;
originally announced March 2019.
-
Multiobjective Optimization in a Quantum Adiabatic Computer
Authors:
Benjamin Baran,
Marcos Villagra
Abstract:
In this work we present a quantum algorithm for multiobjective combinatorial optimization. We show how to map a convex combination of objective functions onto a Hamiltonian and then use that Hamiltonian to prove that the quantum adiabatic algorithm of Farhi \emph{et al.} [arXiv:quant-ph/0001106] can find Pareto-optimal solutions in finite time provided certain convex combinations of objectives are…
▽ More
In this work we present a quantum algorithm for multiobjective combinatorial optimization. We show how to map a convex combination of objective functions onto a Hamiltonian and then use that Hamiltonian to prove that the quantum adiabatic algorithm of Farhi \emph{et al.} [arXiv:quant-ph/0001106] can find Pareto-optimal solutions in finite time provided certain convex combinations of objectives are used and the underlying multiobjective problem meets certain restrictions.
△ Less
Submitted 30 June, 2017; v1 submitted 10 May, 2016;
originally announced May 2016.
-
Dynamic Environments for Virtual Machine Placement considering Elasticity and Overbooking
Authors:
Jammily Ortigoza,
Fabio Lopez-Pires,
Benjamın Baran
Abstract:
Cloud computing datacenters provide millions of virtual machines in actual cloud markets. In this context, Virtual Machine Placement (VMP) is one of the most challenging problems in cloud infrastructure management, considering the large number of possible optimization criteria and different formulations that could be studied. Considering the on-demand model of cloud computing, the VMP problem shou…
▽ More
Cloud computing datacenters provide millions of virtual machines in actual cloud markets. In this context, Virtual Machine Placement (VMP) is one of the most challenging problems in cloud infrastructure management, considering the large number of possible optimization criteria and different formulations that could be studied. Considering the on-demand model of cloud computing, the VMP problem should be solved dynamically to efficiently attend typical workload of modern applications. This work proposes a taxonomy in order to understand possible challenges for Cloud Service Providers (CSPs) in dynamic environments, based on the most relevant dynamic parameters studied so far in the VMP literature. Based on the proposed taxonomy, several unexplored environments have been identified. To further study those research opportunities, sample workload traces for each particular environment are required; therefore, basic examples illustrate a preliminary work on dynamic workload trace generation.
△ Less
Submitted 8 January, 2016;
originally announced January 2016.
-
The XXL Survey: I. Scientific motivations - XMM-Newton observing plan - Follow-up observations and simulation programme
Authors:
M. Pierre,
F. Pacaud,
C. Adami,
S. Alis,
B. Altieri,
B. Baran,
C. Benoist,
M. Birkinshaw,
A. Bongiorno,
M. N. Bremer,
M. Brusa,
A. Butler,
P. Ciliegi,
L. Chiappetti,
N. Clerc,
P. S. Corasaniti,
J. Coupon,
C. De Breuck,
J. Democles,
S. Desai,
J. Delhaize,
J. Devriendt,
Y. Dubois,
D. Eckert,
A. Elyiv
, et al. (67 additional authors not shown)
Abstract:
We present the XXL Survey, the largest XMM programme totaling some 6.9 Ms to date and involving an international consortium of roughly 100 members. The XXL Survey covers two extragalactic areas of 25 deg2 each at a point-source sensitivity of ~ 5E-15 erg/sec/cm2 in the [0.5-2] keV band (completeness limit). The survey's main goals are to provide constraints on the dark energy equation of state fro…
▽ More
We present the XXL Survey, the largest XMM programme totaling some 6.9 Ms to date and involving an international consortium of roughly 100 members. The XXL Survey covers two extragalactic areas of 25 deg2 each at a point-source sensitivity of ~ 5E-15 erg/sec/cm2 in the [0.5-2] keV band (completeness limit). The survey's main goals are to provide constraints on the dark energy equation of state from the space-time distribution of clusters of galaxies and to serve as a pathfinder for future, wide-area X-ray missions. We review science objectives, including cluster studies, AGN evolution, and large-scale structure, that are being conducted with the support of approximately 30 follow-up programmes. We describe the 542 XMM observations along with the associated multi-lambda and numerical simulation programmes. We give a detailed account of the X-ray processing steps and describe innovative tools being developed for the cosmological analysis. The paper provides a thorough evaluation of the X-ray data, including quality controls, photon statistics, exposure and background maps, and sky coverage. Source catalogue construction and multi-lambda associations are briefly described. This material will be the basis for the calculation of the cluster and AGN selection functions, critical elements of the cosmological and science analyses. The XXL multi-lambda data set will have a unique lasting legacy value for cosmological and extragalactic studies and will serve as a calibration resource for future dark energy studies with clusters and other X-ray selected sources. With the present article, we release the XMM XXL photon and smoothed images along with the corresponding exposure maps. The XMM XXL observation list (Table B.1) is available in electronic form at the CDS. The present paper is the first in a series reporting results of the XXL-XMM survey.
△ Less
Submitted 14 December, 2015;
originally announced December 2015.
-
Workload Trace Generation for Dynamic Environments in Cloud Computing
Authors:
Jammily Ortigoza,
Fabio López-Pires,
Benjamín Barán
Abstract:
Cloud computing datacenters provide millions of virtual machines in actual cloud markets. In this context, Virtual Machine Placement (VMP) is one of the most challenging problems in cloud infrastructure management, considering the large number of possible optimization criteria and different formulations that could be studied. Considering the on-demand model of cloud computing, the VMP problem shou…
▽ More
Cloud computing datacenters provide millions of virtual machines in actual cloud markets. In this context, Virtual Machine Placement (VMP) is one of the most challenging problems in cloud infrastructure management, considering the large number of possible optimization criteria and different formulations that could be studied. Considering the on-demand model of cloud computing, the VMP problem should be optimized dynamically to efficiently attend typical workload of modern applications. This work proposes several dynamic environments for solving the VMP from the providers' perspective based on the most relevant dynamic parameters studied so far in the VMP literature. A complete set of environments and workload traces examples are presented in this work.
△ Less
Submitted 7 November, 2015; v1 submitted 30 June, 2015;
originally announced July 2015.
-
Virtual Machine Placement Literature Review
Authors:
Fabio Lopez-Pires,
Benjamin Baran
Abstract:
Cloud Computing Datacenters host millions of virtual machines (VMs) on real world scenarios. In this context, Virtual Machine Placement (VMP) is one of the most challenging problems in cloud infrastructure management, considering also the large number of possible optimization criteria and different formulations that could be studied. VMP literature include relevant topics such as energy-efficiency…
▽ More
Cloud Computing Datacenters host millions of virtual machines (VMs) on real world scenarios. In this context, Virtual Machine Placement (VMP) is one of the most challenging problems in cloud infrastructure management, considering also the large number of possible optimization criteria and different formulations that could be studied. VMP literature include relevant topics such as energy-efficiency, Service Level Agreements (SLA), cloud service markets, Quality of Service (QoS) and carbon dioxide emissions, all of them with high economical and ecological impact. This work presents an extensive up-to-date review of the most relevant VMP literature in order to identify research opportunities.
△ Less
Submitted 4 June, 2015;
originally announced June 2015.
-
Fitting ideals and the Gorenstein property
Authors:
Burcu Baran
Abstract:
Let p be a prime number and G be a finite commutative group such that p^{2} does not divide the order of G. In this note we prove that for every finite module M over the group ring Z_{p}[G], the inequality #M \leq #Z_{p}[G]/Fit_{Z_{p}[G]}(M) holds. Here, Fit_{Z_{p}[G]}(M) is the Z_{p}[G]-Fitting ideal of M.
Let p be a prime number and G be a finite commutative group such that p^{2} does not divide the order of G. In this note we prove that for every finite module M over the group ring Z_{p}[G], the inequality #M \leq #Z_{p}[G]/Fit_{Z_{p}[G]}(M) holds. Here, Fit_{Z_{p}[G]}(M) is the Z_{p}[G]-Fitting ideal of M.
△ Less
Submitted 14 October, 2009; v1 submitted 18 February, 2009;
originally announced February 2009.
-
A Modular Curve of Level 9 and the Class Number One Problem
Authors:
Burcu Baran
Abstract:
In this note we give an explicit parametrization of the modular curve associated to the normalizer of a non-split Cartan subgroup of level 9. We determine all integral points of this modular curve. As an application, we give an alternative solution to the class number one problem.
In this note we give an explicit parametrization of the modular curve associated to the normalizer of a non-split Cartan subgroup of level 9. We determine all integral points of this modular curve. As an application, we give an alternative solution to the class number one problem.
△ Less
Submitted 18 February, 2009; v1 submitted 30 January, 2008;
originally announced January 2008.
