-
Diversified in-domain synthesis with efficient fine-tuning for few-shot classification
Authors:
Victor G. Turrisi da Costa,
Nicola Dall'Asen,
Yiming Wang,
Nicu Sebe,
Elisa Ricci
Abstract:
Few-shot image classification aims to learn an image classifier using only a small set of labeled examples per class. A recent research direction for improving few-shot classifiers involves augmenting the labelled samples with synthetic images created by state-of-the-art text-to-image generation models. Following this trend, we propose Diversified In-domain Synthesis with Efficient Fine-tuning (DI…
▽ More
Few-shot image classification aims to learn an image classifier using only a small set of labeled examples per class. A recent research direction for improving few-shot classifiers involves augmenting the labelled samples with synthetic images created by state-of-the-art text-to-image generation models. Following this trend, we propose Diversified In-domain Synthesis with Efficient Fine-tuning (DISEF), a novel approach which addresses the generalization challenge in few-shot learning using synthetic data. DISEF consists of two main components. First, we propose a novel text-to-image augmentation pipeline that, by leveraging the real samples and their rich semantics coming from an advanced captioning model, promotes in-domain sample diversity for better generalization. Second, we emphasize the importance of effective model fine-tuning in few-shot recognition, proposing to use Low-Rank Adaptation (LoRA) for joint adaptation of the text and image encoders in a Vision Language Model. We validate our method in ten different benchmarks, consistently outperforming baselines and establishing a new state-of-the-art for few-shot classification. Code is available at https://github.com/vturrisi/disef.
△ Less
Submitted 6 December, 2023; v1 submitted 5 December, 2023;
originally announced December 2023.
-
Encoding information onto the charge and spin state of a paramagnetic atom using MgO tunnelling spintronics
Authors:
Mathieu Lamblin,
Bhavishya Chowrira,
Victor Da Costa,
Bertrand Vileno,
Loic Joly,
Samy Boukari,
Wolfgang Weber,
Romain Bernard,
Benoit Gobaut,
Michel Hehn,
Daniel Lacour,
Martin Bowen
Abstract:
An electrical current that flows across individual atoms or molecules can generate exotic quantum-based behavior, from memristive effects to Coulomb blockade and the promotion of quantum excited states. These fundamental effects typically appear one at a time in model junctions built using atomic tip or lateral techniques. So far, however, a viable industrial pathway for such discrete state device…
▽ More
An electrical current that flows across individual atoms or molecules can generate exotic quantum-based behavior, from memristive effects to Coulomb blockade and the promotion of quantum excited states. These fundamental effects typically appear one at a time in model junctions built using atomic tip or lateral techniques. So far, however, a viable industrial pathway for such discrete state devices has been lacking. Here, we demonstrate that a commercialized device platform can serve as this industrial pathway for quantum technologies. We have studied magnetic tunnel junctions with a MgO barrier containing C atoms. The paramagnetic localized electrons due to individual C atoms generate parallel nanotransport paths across the micronic device as deduced from magnetotransport experiments. Coulomb blockade effects linked to tunnelling magnetoresistance peaks can be electrically controlled, leading to a persistent memory effect. Our results position MgO tunneling spintronics as a promising platform to industrially implement quantum technologies.
△ Less
Submitted 31 August, 2023;
originally announced August 2023.
-
Simplifying Open-Set Video Domain Adaptation with Contrastive Learning
Authors:
Giacomo Zara,
Victor Guilherme Turrisi da Costa,
Subhankar Roy,
Paolo Rota,
Elisa Ricci
Abstract:
In an effort to reduce annotation costs in action recognition, unsupervised video domain adaptation methods have been proposed that aim to adapt a predictive model from a labelled dataset (i.e., source domain) to an unlabelled dataset (i.e., target domain). In this work we address a more realistic scenario, called open-set video domain adaptation (OUVDA), where the target dataset contains "unknown…
▽ More
In an effort to reduce annotation costs in action recognition, unsupervised video domain adaptation methods have been proposed that aim to adapt a predictive model from a labelled dataset (i.e., source domain) to an unlabelled dataset (i.e., target domain). In this work we address a more realistic scenario, called open-set video domain adaptation (OUVDA), where the target dataset contains "unknown" semantic categories that are not shared with the source. The challenge lies in aligning the shared classes of the two domains while separating the shared classes from the unknown ones. In this work we propose to address OUVDA with an unified contrastive learning framework that learns discriminative and well-clustered features. We also propose a video-oriented temporal contrastive loss that enables our method to better cluster the feature space by exploiting the freely available temporal information in video data. We show that discriminative feature space facilitates better separation of the unknown classes, and thereby allows us to use a simple similarity based score to identify them. We conduct thorough experimental evaluation on multiple OUVDA benchmarks and show the effectiveness of our proposed method against the prior art.
△ Less
Submitted 9 January, 2023;
originally announced January 2023.
-
Bayesian Prompt Learning for Image-Language Model Generalization
Authors:
Mohammad Mahdi Derakhshani,
Enrique Sanchez,
Adrian Bulat,
Victor Guilherme Turrisi da Costa,
Cees G. M. Snoek,
Georgios Tzimiropoulos,
Brais Martinez
Abstract:
Foundational image-language models have generated considerable interest due to their efficient adaptation to downstream tasks by prompt learning. Prompt learning treats part of the language model input as trainable while freezing the rest, and optimizes an Empirical Risk Minimization objective. However, Empirical Risk Minimization is known to suffer from distributional shifts which hurt generaliza…
▽ More
Foundational image-language models have generated considerable interest due to their efficient adaptation to downstream tasks by prompt learning. Prompt learning treats part of the language model input as trainable while freezing the rest, and optimizes an Empirical Risk Minimization objective. However, Empirical Risk Minimization is known to suffer from distributional shifts which hurt generalizability to prompts unseen during training. By leveraging the regularization ability of Bayesian methods, we frame prompt learning from the Bayesian perspective and formulate it as a variational inference problem. Our approach regularizes the prompt space, reduces overfitting to the seen prompts and improves the prompt generalization on unseen prompts. Our framework is implemented by modeling the input prompt space in a probabilistic manner, as an a priori distribution which makes our proposal compatible with prompt learning approaches that are unconditional or conditional on the image. We demonstrate empirically on 15 benchmarks that Bayesian prompt learning provides an appropriate coverage of the prompt space, prevents learning spurious features, and exploits transferable invariant features. This results in better generalization of unseen prompts, even across different datasets and domains. Code available at: https://github.com/saic-fi/Bayesian-Prompt-Learning
△ Less
Submitted 20 August, 2023; v1 submitted 5 October, 2022;
originally announced October 2022.
-
Unsupervised Domain Adaptation for Video Transformers in Action Recognition
Authors:
Victor G. Turrisi da Costa,
Giacomo Zara,
Paolo Rota,
Thiago Oliveira-Santos,
Nicu Sebe,
Vittorio Murino,
Elisa Ricci
Abstract:
Over the last few years, Unsupervised Domain Adaptation (UDA) techniques have acquired remarkable importance and popularity in computer vision. However, when compared to the extensive literature available for images, the field of videos is still relatively unexplored. On the other hand, the performance of a model in action recognition is heavily affected by domain shift. In this paper, we propose…
▽ More
Over the last few years, Unsupervised Domain Adaptation (UDA) techniques have acquired remarkable importance and popularity in computer vision. However, when compared to the extensive literature available for images, the field of videos is still relatively unexplored. On the other hand, the performance of a model in action recognition is heavily affected by domain shift. In this paper, we propose a simple and novel UDA approach for video action recognition. Our approach leverages recent advances on spatio-temporal transformers to build a robust source model that better generalises to the target domain. Furthermore, our architecture learns domain invariant features thanks to the introduction of a novel alignment loss term derived from the Information Bottleneck principle. We report results on two video action recognition benchmarks for UDA, showing state-of-the-art performance on HMDB$\leftrightarrow$UCF, as well as on Kinetics$\rightarrow$NEC-Drone, which is more challenging. This demonstrates the effectiveness of our method in handling different levels of domain shift. The source code is available at https://github.com/vturrisi/UDAVT.
△ Less
Submitted 26 July, 2022;
originally announced July 2022.
-
A Gaseous Argon-Based Near Detector to Enhance the Physics Capabilities of DUNE
Authors:
A. Abed Abud,
B. Abi,
R. Acciarri,
M. A. Acero,
M. R. Adames,
G. Adamov,
M. Adamowski,
D. Adams,
M. Adinolfi,
C. Adriano,
A. Aduszkiewicz,
J. Aguilar,
Z. Ahmad,
J. Ahmed,
B. Aimard,
F. Akbar,
B. Ali-Mohammadzadeh,
T. Alion,
K. Allison,
S. Alonso Monsalve,
M. AlRashed,
C. Alt,
A. Alton,
R. Alvarez,
P. Amedo
, et al. (1220 additional authors not shown)
Abstract:
This document presents the concept and physics case for a magnetized gaseous argon-based detector system (ND-GAr) for the Deep Underground Neutrino Experiment (DUNE) Near Detector. This detector system is required in order for DUNE to reach its full physics potential in the measurement of CP violation and in delivering precision measurements of oscillation parameters. In addition to its critical r…
▽ More
This document presents the concept and physics case for a magnetized gaseous argon-based detector system (ND-GAr) for the Deep Underground Neutrino Experiment (DUNE) Near Detector. This detector system is required in order for DUNE to reach its full physics potential in the measurement of CP violation and in delivering precision measurements of oscillation parameters. In addition to its critical role in the long-baseline oscillation program, ND-GAr will extend the overall physics program of DUNE. The LBNF high-intensity proton beam will provide a large flux of neutrinos that is sampled by ND-GAr, enabling DUNE to discover new particles and search for new interactions and symmetries beyond those predicted in the Standard Model.
△ Less
Submitted 11 March, 2022;
originally announced March 2022.
-
Toward accurate polarization estimation in nanoscopic systems
Authors:
Sambit Mohapatra,
Wolfgang Weber,
Martin Bowen,
Samy Boukari,
Victor Da Costa
Abstract:
The nanoscopic characterization of ferroelectric thin films is crucial from their device application point of view. Standard characterization techniques are based on detecting the nanoscopic charge compensation current (switching current) caused by the polarization reversal in the ferroelectric. Owing to various surface and bulk limited mechanisms, leakage currents commonly appear during such meas…
▽ More
The nanoscopic characterization of ferroelectric thin films is crucial from their device application point of view. Standard characterization techniques are based on detecting the nanoscopic charge compensation current (switching current) caused by the polarization reversal in the ferroelectric. Owing to various surface and bulk limited mechanisms, leakage currents commonly appear during such measurements, which are frequently subtracted using the device I-V characteristic by employing positive-up-negative-down (PUND) measurement scheme. By performing nanoscopic switching current measurements on a commonly used ferroelectric, BiFeO3, we show that such characterization methods may be prone to large errors in the polarization estimation on ferro-resistive samples, due to current background subtraction issues. Especially, when ferro-resistive behavior is associated with the polarization reversal of the ferroelectric thin film, background current subtraction is not accurate due to the mismatch of the I-V characteristics for the two polarization states. We show instead that removing the background current by an asymmetric least squares subtraction method, though not perfect, gives a much better estimation of the ferroelectric properties of the sample under study.
△ Less
Submitted 11 March, 2022;
originally announced March 2022.
-
Snowmass Neutrino Frontier: DUNE Physics Summary
Authors:
DUNE Collaboration,
A. Abed Abud,
B. Abi,
R. Acciarri,
M. A. Acero,
M. R. Adames,
G. Adamov,
M. Adamowski,
D. Adams,
M. Adinolfi,
C. Adriano,
A. Aduszkiewicz,
J. Aguilar,
Z. Ahmad,
J. Ahmed,
B. Aimard,
F. Akbar,
B. Ali-Mohammadzadeh,
T. Alion,
K. Allison,
S. Alonso Monsalve,
M. AlRashed,
C. Alt,
A. Alton,
R. Alvarez
, et al. (1221 additional authors not shown)
Abstract:
The Deep Underground Neutrino Experiment (DUNE) is a next-generation long-baseline neutrino oscillation experiment with a primary physics goal of observing neutrino and antineutrino oscillation patterns to precisely measure the parameters governing long-baseline neutrino oscillation in a single experiment, and to test the three-flavor paradigm. DUNE's design has been developed by a large, internat…
▽ More
The Deep Underground Neutrino Experiment (DUNE) is a next-generation long-baseline neutrino oscillation experiment with a primary physics goal of observing neutrino and antineutrino oscillation patterns to precisely measure the parameters governing long-baseline neutrino oscillation in a single experiment, and to test the three-flavor paradigm. DUNE's design has been developed by a large, international collaboration of scientists and engineers to have unique capability to measure neutrino oscillation as a function of energy in a broadband beam, to resolve degeneracy among oscillation parameters, and to control systematic uncertainty using the exquisite imaging capability of massive LArTPC far detector modules and an argon-based near detector. DUNE's neutrino oscillation measurements will unambiguously resolve the neutrino mass ordering and provide the sensitivity to discover CP violation in neutrinos for a wide range of possible values of $δ_{CP}$. DUNE is also uniquely sensitive to electron neutrinos from a galactic supernova burst, and to a broad range of physics beyond the Standard Model (BSM), including nucleon decays. DUNE is anticipated to begin collecting physics data with Phase I, an initial experiment configuration consisting of two far detector modules and a minimal suite of near detector components, with a 1.2 MW proton beam. To realize its extensive, world-leading physics potential requires the full scope of DUNE be completed in Phase II. The three Phase II upgrades are all necessary to achieve DUNE's physics goals: (1) addition of far detector modules three and four for a total FD fiducial mass of at least 40 kt, (2) upgrade of the proton beam power from 1.2 MW to 2.4 MW, and (3) replacement of the near detector's temporary muon spectrometer with a magnetized, high-pressure gaseous argon TPC and calorimeter.
△ Less
Submitted 11 March, 2022;
originally announced March 2022.
-
Using the Energy probability distribution zeros to obtain the critical properties of the two-dimensional anisotropic Heisenberg model
Authors:
Gabriel Bruno Garcia de Souza,
Bismarck Vaz da Costa
Abstract:
In this paper we present a Monte Carlo study of the critical behavior of the easy axis anisotropic Heisenberg spin model in two dimensions. Based on the partial knowledge of the zeros of the energy probability distribution we determine with good precision the phase diagram of the model obtaining the critical temperature and exponents for several values of the anisotropy. Our results indicate that…
▽ More
In this paper we present a Monte Carlo study of the critical behavior of the easy axis anisotropic Heisenberg spin model in two dimensions. Based on the partial knowledge of the zeros of the energy probability distribution we determine with good precision the phase diagram of the model obtaining the critical temperature and exponents for several values of the anisotropy. Our results indicate that the model is in the Ising universality class for any anisotropy.
△ Less
Submitted 7 July, 2022; v1 submitted 1 March, 2022;
originally announced March 2022.
-
Self-Supervised Models are Continual Learners
Authors:
Enrico Fini,
Victor G. Turrisi da Costa,
Xavier Alameda-Pineda,
Elisa Ricci,
Karteek Alahari,
Julien Mairal
Abstract:
Self-supervised models have been shown to produce comparable or better visual representations than their supervised counterparts when trained offline on unlabeled data at scale. However, their efficacy is catastrophically reduced in a Continual Learning (CL) scenario where data is presented to the model sequentially. In this paper, we show that self-supervised loss functions can be seamlessly conv…
▽ More
Self-supervised models have been shown to produce comparable or better visual representations than their supervised counterparts when trained offline on unlabeled data at scale. However, their efficacy is catastrophically reduced in a Continual Learning (CL) scenario where data is presented to the model sequentially. In this paper, we show that self-supervised loss functions can be seamlessly converted into distillation mechanisms for CL by adding a predictor network that maps the current state of the representations to their past state. This enables us to devise a framework for Continual self-supervised visual representation Learning that (i) significantly improves the quality of the learned representations, (ii) is compatible with several state-of-the-art self-supervised objectives, and (iii) needs little to no hyperparameter tuning. We demonstrate the effectiveness of our approach empirically by training six popular self-supervised models in various CL settings.
△ Less
Submitted 1 April, 2022; v1 submitted 8 December, 2021;
originally announced December 2021.
-
Organic ferroelectric Croconic Acid: A concise survey from bulk single crystals to thin films
Authors:
Sambit Mohapatra,
Salia Cherifi-Hertel,
Senthil Kumar Kuppusamy,
Guy Schmerber,
Jacek Arabski,
Benoit Gobaut,
Wolfgang Weber,
Martin Bowen,
Victor Da Costa,
Samy Boukari
Abstract:
Owing to prospective energy-efficient and environmentally benign applications, organic ferroelectric materials are useful and necessary alternative to inorganic ferroelectrics. Although the first discovered ferroelectric, Rochelle salt, was a salt of an organic compound, organic ferroelectrics have not been as abundant as the inorganic ones. Further, the small polarization values in the organic sy…
▽ More
Owing to prospective energy-efficient and environmentally benign applications, organic ferroelectric materials are useful and necessary alternative to inorganic ferroelectrics. Although the first discovered ferroelectric, Rochelle salt, was a salt of an organic compound, organic ferroelectrics have not been as abundant as the inorganic ones. Further, the small polarization values in the organic systems discovered so far have been a demotivating factor for their applications. However, scientific interest and activities surrounding such materials, for the purpose of fundamental understanding and practical applications, have significantly risen lately, especially after the discovery of above-room-temperature ferroelectricity in croconic acid (4,5-dihydroxy-4-cyclopentene-1,2,3-trione, H2C5O5) crystals with polarization values rivalling those found in inorganic ferroelectrics. Its large polarization, organic nature, and vacuum sublimability make croconic acid an ideal candidate for non-toxic and lead-free device applications. In this review article, we survey the scientific activities carried out so far involving ferroelectricity in this novel material, paying equal attention to its bulk single crystal and thin film forms. While we discuss about the origin of ferroelectric order and the reversal of polarization in the bulk form, we also summarize the directions toward applications of the thin films.
△ Less
Submitted 24 October, 2021;
originally announced October 2021.
-
Solo-learn: A Library of Self-supervised Methods for Visual Representation Learning
Authors:
Victor G. Turrisi da Costa,
Enrico Fini,
Moin Nabi,
Nicu Sebe,
Elisa Ricci
Abstract:
This paper presents solo-learn, a library of self-supervised methods for visual representation learning. Implemented in Python, using Pytorch and Pytorch lightning, the library fits both research and industry needs by featuring distributed training pipelines with mixed-precision, faster data loading via Nvidia DALI, online linear evaluation for better prototyping, and many additional training tric…
▽ More
This paper presents solo-learn, a library of self-supervised methods for visual representation learning. Implemented in Python, using Pytorch and Pytorch lightning, the library fits both research and industry needs by featuring distributed training pipelines with mixed-precision, faster data loading via Nvidia DALI, online linear evaluation for better prototyping, and many additional training tricks. Our goal is to provide an easy-to-use library comprising a large amount of Self-supervised Learning (SSL) methods, that can be easily extended and fine-tuned by the community. solo-learn opens up avenues for exploiting large-budget SSL solutions on inexpensive smaller infrastructures and seeks to democratize SSL by making it accessible to all. The source code is available at https://github.com/vturrisi/solo-learn.
△ Less
Submitted 4 February, 2022; v1 submitted 3 August, 2021;
originally announced August 2021.
-
Nanoscale reversal of stable room temperature ferroelectric polarization in organic croconic acid thin films
Authors:
Sambit Mohapatra,
Eric Beaurepaire,
Wolfgang Weber,
Martin Bowen,
Samy Boukari,
Victor Da Costa
Abstract:
It was discovered in 2010 that Croconic Acid, in its crystal form, has the highest polarization among organic ferroelectrics. In the context of eliminating toxic substances from electronic devices, Croconic Acid has a great potential as a sublimable lead-free ferroelectric. However, studies on ferroelectric properties of its thin films are only in their early stages and its capability to be incorp…
▽ More
It was discovered in 2010 that Croconic Acid, in its crystal form, has the highest polarization among organic ferroelectrics. In the context of eliminating toxic substances from electronic devices, Croconic Acid has a great potential as a sublimable lead-free ferroelectric. However, studies on ferroelectric properties of its thin films are only in their early stages and its capability to be incorporated in nanoscale devices is unknown. In this work, we demonstrate, upon ferroelectric switching at the nanoscale, stable and enduring room temperature polarization with no leakage current in Croconic Acid thin films. We thus show that it is a promising lead-free organic ferroelectric toward integration in nanoscale devices. The challenging switching current and polarization reversal characterization at the nanoscale was done using a unique combination of piezoresponse force microscopy, polarization switching current spectroscopy and the concurrent electromechanical strain response. Indeed, this combination can help to rationalize otherwise asymmetric polarization-voltage data and distorted hysteresis due to current jumps below the background noise, which are statistically washed away in macrojunctions but become prevalent at the nanoscale. These results are valid irrespective of the ferroelectrics' nature, organic or inorganic. Beyond the potential of Croconic Acid as an ecological ferroelectric material in devices, our detection of a clear nanoscopic polarization switching current thus paves the way for a fundamental understanding and technological applications of the polarization reversal mechanism at the nanoscale.
△ Less
Submitted 7 April, 2021;
originally announced April 2021.
-
Quantum advantage in a molecular spintronic engine that harvests thermal fluctuation energy
Authors:
Bhavishya Chowrira,
Lalit Kandpal,
Mathieu Lamblin,
Franck Ngassam,
Charles-Ambroise Kouakou,
Talha Zafar,
Damien Mertz,
Bertrand Vileno,
Christophe Kieber,
Gilles Versini,
Benoit Gobaut,
Loic Joly,
Tom Ferte,
Elmer Monteblanco,
Armel Bahouka,
Romain Bernard,
Sambit Mohapatra,
H. Prima Garcia,
S. Elidrissi,
M. Gavara,
Emmanuel Sternitzky,
Victor Da Costa,
Michel Hehn,
Francois Montaigne,
Fadi Choueikani
, et al. (6 additional authors not shown)
Abstract:
Recent theory and experiments have showcased how to harness quantum mechanics to assemble heat/information engines with efficiencies that surpass the classical Carnot limit. So far, this has required atomic engines that are driven by cumbersome external electromagnetic sources. Here, using molecular spintronics, we propose an implementation that is both electronic and autonomous. Our spintronic qu…
▽ More
Recent theory and experiments have showcased how to harness quantum mechanics to assemble heat/information engines with efficiencies that surpass the classical Carnot limit. So far, this has required atomic engines that are driven by cumbersome external electromagnetic sources. Here, using molecular spintronics, we propose an implementation that is both electronic and autonomous. Our spintronic quantum engine heuristically deploys several known quantum assets by having a chain of spin qubits formed by the paramagnetic Co centers of phthalocyanine (Pc) molecules electronically interact with electron-spin selecting Fe/C60 interfaces. Density functional calculations reveal that transport fluctuations across the interface can stabilize spin coherence on the Co paramagnetic centers, which host spin flip processes. Across vertical molecular nanodevices, we measure enduring dc current generation, output power above room temperature, two quantum thermodynamical signatures of the engine's processes, and a record 89% spin polarization of current across the Fe/C60 interface. It is crucially this electron spin selection that forces, through demonic feedback and control, charge current to flow against the built-in potential barrier. Further research into spintronic quantum engines, insight into the quantum information processes within spintronic technologies, and retooling the spintronic-based information technology chain, could help accelerate the transition to clean energy.
△ Less
Submitted 18 August, 2022; v1 submitted 22 September, 2020;
originally announced September 2020.
-
A Parametrized Nonlinear Predictive Control Strategy for Relaxing COVID-19 Social Distancing Measures in Brazil
Authors:
Marcelo M. Morato,
Igor M. L. Pataro,
Marcus V. Americano da Costa,
Julio E. Normey-Rico
Abstract:
In this paper, we formulate a Nonlinear Model Predictive Control (NMPC) to plan appropriate social distancing measures (and relaxations) in order to mitigate the COVID-19 pandemic effects, considering the contagion development in Brazil. The NMPC strategy is designed upon an adapted data-driven Susceptible-Infected-Recovered-Deceased (SIRD) contagion model, which takes into account the effects of…
▽ More
In this paper, we formulate a Nonlinear Model Predictive Control (NMPC) to plan appropriate social distancing measures (and relaxations) in order to mitigate the COVID-19 pandemic effects, considering the contagion development in Brazil. The NMPC strategy is designed upon an adapted data-driven Susceptible-Infected-Recovered-Deceased (SIRD) contagion model, which takes into account the effects of social distancing. Furthermore, the adapted SIRD model includes time-varying auto-regressive contagion parameters, which dynamically converge according to the stage of the pandemic. This new model is identified through a three-layered procedures, with analytical regressions, Least-Squares optimization runs and auto-regressive model fits. The data-driven model is validated and shown to adequately describe the contagion curves over large forecast horizons. In this model, control input is defined as finitely parametrized values for social distancing guidelines, which directly affect the transmission and infection rates of the SARS-CoV-2 virus. The NMPC strategy generates piece-wise constant quarantine guidelines which can be relaxed/strengthen as each week passes. The implementation of the method is pursued through a search mechanism, since the control is finitely parametrized and, thus, there exist a finite number of possible control sequences. Simulation essays are shown to illustrate the results obtained with the proposed closed-loop NMPC strategy, which is able to mitigate the number of infections and progressively loosen social distancing measures. With respect to an "open-loop"/no control condition, the number of deaths still could be reduced in up to 30 %. The forecast preview an infection peak to September 2nd, 2020, which could lead to over 1.5 million deaths if no coordinate health policy is enacted. The framework serves as guidelines for possible public health policies in Brazil.
△ Less
Submitted 19 July, 2020;
originally announced July 2020.
-
Optimal Control Concerns Regarding the COVID-19 (SARS-CoV-2) Pandemic in Bahia and Santa Catarina, Brazil
Authors:
Marcelo M. Morato,
Igor M. L. Pataro,
Marcus V. Americano da Costa,
Julio E. Normey-Rico
Abstract:
The COVID-19 pandemic is the profoundest health crisis of the 21rst century. The SARS-CoV-2 virus arrived in Brazil around March, 2020 and its social and economical backlashes are catastrophic. In this paper, it is investigated how Model Predictive Control (MPC) could be used to plan appropriate social distancing policies to mitigate the pandemic effects in Bahia and Santa Catarina, two states of…
▽ More
The COVID-19 pandemic is the profoundest health crisis of the 21rst century. The SARS-CoV-2 virus arrived in Brazil around March, 2020 and its social and economical backlashes are catastrophic. In this paper, it is investigated how Model Predictive Control (MPC) could be used to plan appropriate social distancing policies to mitigate the pandemic effects in Bahia and Santa Catarina, two states of different regions, culture, and population demography in Brazil. In addition, the parameters of Susceptible-Infected-Recovered-Deceased (SIRD) models for these two states are identified using an optimization procedure. The control input to the process is a social isolation guideline passed to the population. Two MPC strategies are designed: a) a centralized MPC, which coordinates a single control policy for both states; and b) a decentralized strategy, for which one optimization is solved for each state. Simulation results are shown to illustrate and compare both control strategies. The framework serves as guidelines to deals with such pandemic phenomena.
△ Less
Submitted 24 June, 2020;
originally announced June 2020.
-
Vortices in Kekulene Molecules
Authors:
Lucas Menicucci,
Francisco César Sá Barreto,
Bismarck vaz da Costa
Abstract:
Kekulene is an aromatic hydrocarbon with formula C48H24 arranged in the shape of a closed super-ring as shown in Fig. 2. It consists of a sublattice with 48 C atoms with spin 5/2 and a 24 hydrogen sublattice with spin 2. In this communication, we use Monte Carlo simulations to determine the magnetic structures present in Kekulene for several temperatures (T) and dipole anisotropies (δ = D/J). Our…
▽ More
Kekulene is an aromatic hydrocarbon with formula C48H24 arranged in the shape of a closed super-ring as shown in Fig. 2. It consists of a sublattice with 48 C atoms with spin 5/2 and a 24 hydrogen sublattice with spin 2. In this communication, we use Monte Carlo simulations to determine the magnetic structures present in Kekulene for several temperatures (T) and dipole anisotropies (δ = D/J). Our results show that there are two regimes at low temperature separated by a crossover at 2.5 < δcross < 3.0. For δ < δcross the ground state has a unique vortex configuration. In the region δ > δcross arrangements of vortices-antivortices (V-AV) appears. As temperature raises the vortex structure disorders and small oscillations take over. The importance on synthesizing this molecule grounds in the possibility of building real planar structures of sizes at least 10 times smaller than the earlier proposed permalloy nanodots. It is worthy to mention that Kekulene is a planar structure with atomic thickness, which is a great advantage compared with other nanomagnetic structures.
△ Less
Submitted 22 June, 2020;
originally announced June 2020.
-
Robust ferroelectric properties of organic Croconic Acid films grown on spintronically relevant substrates
Authors:
Sambit Mohapatra,
Victor Da Costa,
Garen Avedissian,
Jacek Arabski,
Wolfgang Weber,
Martin Bowen,
Samy Boukari
Abstract:
The discovery of stable room temperature ferroelectricity in Croconic Acid, an organic ferroelectric material, with polarization values on par with those found in inorganic ferroelectric materials and highest among organic ferroelectric materials, has opened up possibilities to realize myriads of nano-electronic and spintronic devices based on organic ferroelectrics. Such possibilities require an…
▽ More
The discovery of stable room temperature ferroelectricity in Croconic Acid, an organic ferroelectric material, with polarization values on par with those found in inorganic ferroelectric materials and highest among organic ferroelectric materials, has opened up possibilities to realize myriads of nano-electronic and spintronic devices based on organic ferroelectrics. Such possibilities require an adequate understanding of the ferroelectric properties of Croconic Acid grown on surfaces that are commonly employed in device fabrication. While several macroscopic studies on relatively larger crystals of Croconic Acid have been performed, studies on thin films are only in their early stages. We have grown thin films of Croconic Acid on gold and cobalt surfaces, which are commonly used in spintronic devices as metallic electrodes, and studied the ferroelectric response of the films using ex-situ Piezoresponse Force Microscopy at room temperature. We show that the polarization reversal in Croconic Acid domains is sensitive to the substrate surface. Using the same experimental protocol, we observe the robust polarization reversal of a single, mostly in-plane electrical domain for a cobalt substrate, whereas polarization reversal is hardly observed for a gold substrate. We attribute this difference to the influence of substrates on the Croconic Acid molecular networks. Our study suggests that to realize devices one has to take care about the substrate on which Croconic Acid will be deposited. The fact that polarization switching is robust on cobalt surface can be used to fabricate multifunctional devices that utilize the cobalt-Croconic Acid interface.
△ Less
Submitted 9 April, 2020;
originally announced April 2020.
-
Magnetoresistance and spintronic anisotropy induced by spin excitations along molecular spin chains
Authors:
K. Katcko,
E. Urbain,
L. Kandpal,
B. Chowrira,
F. Schleicher,
U. Halisdemir,
F. Ngassamnyakam,
D. Mertz,
B. Leconte,
N. Beyer,
D. Spor,
P. Panissod,
A. Boulard,
J. Arabski,
C. Kieber,
E. Sternitsky,
V. Da Costa,
M. Alouani,
M. Hehn,
F. Montaigne,
A. Bahouka,
W. Weber,
E. Beaurepaire,
D. Lacour,
S. Boukari
, et al. (1 additional authors not shown)
Abstract:
Electrically manipulating the quantum properties of nano-objects, such as atoms or molecules, is typically done using scanning tunnelling microscopes and lateral junctions. The resulting nanotransport path is well established in these model devices. Societal applications require transposing this knowledge to nano-objects embedded within vertical solid-state junctions, which can advantageously harn…
▽ More
Electrically manipulating the quantum properties of nano-objects, such as atoms or molecules, is typically done using scanning tunnelling microscopes and lateral junctions. The resulting nanotransport path is well established in these model devices. Societal applications require transposing this knowledge to nano-objects embedded within vertical solid-state junctions, which can advantageously harness spintronics to address these quantum properties thanks to ferromagnetic electrodes and high-quality interfaces. The challenge here is to ascertain the device's effective, buried nanotransport path, and to electrically involve these nano-objects in this path by shrinking the device area from the macro- to the nano-scale while maintaining high structural/chemical quality across the heterostructure. We've developed a low-tech, resist- and solvent-free technological process that can craft nanopillar devices from entire in-situ grown heterostructures, and use it to study magnetotransport between two Fe and Co ferromagnetic electrodes across a functional magnetic CoPc molecular layer. We observe how spin-flip transport across CoPc molecular spin chains promotes a specific magnetoresistance effect, and alters the nanojunction's magnetism through spintronic anisotropy. In the process, we identify three magnetic units along the effective nanotransport path thanks to a macrospin model of magnetotransport. Our work elegantly connects the until now loosely associated concepts of spin-flip spectroscopy, magnetic exchange bias and magnetotransport due to molecular spin chains, within a solid-state device. We notably measure a 5.9meV energy threshold for magnetic decoupling between the Fe layer's buried atoms and those in contact with the CoPc layer forming the so-called 'spinterface'. This provides a first insight into the experimental energetics of this promising low-power information encoding unit.
△ Less
Submitted 22 November, 2019; v1 submitted 23 October, 2019;
originally announced October 2019.
-
Online Local Boosting: improving performance in online decision trees
Authors:
Victor G. Turrisi da Costa,
Saulo Martiello Mastelini,
André C. Ponce de Leon Ferreira de Carvalho,
Sylvio Barbon Jr
Abstract:
As more data are produced each day, and faster, data stream mining is growing in importance, making clear the need for algorithms able to fast process these data. Data stream mining algorithms are meant to be solutions to extract knowledge online, specially tailored from continuous data problem. Many of the current algorithms for data stream mining have high processing and memory costs. Often, the…
▽ More
As more data are produced each day, and faster, data stream mining is growing in importance, making clear the need for algorithms able to fast process these data. Data stream mining algorithms are meant to be solutions to extract knowledge online, specially tailored from continuous data problem. Many of the current algorithms for data stream mining have high processing and memory costs. Often, the higher the predictive performance, the higher these costs. To increase predictive performance without largely increasing memory and time costs, this paper introduces a novel algorithm, named Online Local Boosting (OLBoost), which can be combined into online decision tree algorithms to improve their predictive performance without modifying the structure of the induced decision trees. For such, OLBoost applies a boosting to small separate regions of the instances space. Experimental results presented in this paper show that by using OLBoost the online learning decision tree algorithms can significantly improve their predictive performance. Additionally, it can make smaller trees perform as good or better than larger trees.
△ Less
Submitted 16 July, 2019;
originally announced July 2019.
-
Strict Very Fast Decision Tree: a memory conservative algorithm for data stream mining
Authors:
Victor Guilherme Turrisi da Costa,
André Carlos Ponce de Leon Ferreira de Carvalho,
Sylvio Barbon Junior
Abstract:
Dealing with memory and time constraints are current challenges when learning from data streams with a massive amount of data. Many algorithms have been proposed to handle these difficulties, among them, the Very Fast Decision Tree (VFDT) algorithm. Although the VFDT has been widely used in data stream mining, in the last years, several authors have suggested modifications to increase its performa…
▽ More
Dealing with memory and time constraints are current challenges when learning from data streams with a massive amount of data. Many algorithms have been proposed to handle these difficulties, among them, the Very Fast Decision Tree (VFDT) algorithm. Although the VFDT has been widely used in data stream mining, in the last years, several authors have suggested modifications to increase its performance, putting aside memory concerns by proposing memory-costly solutions. Besides, most data stream mining solutions have been centred around ensembles, which combine the memory costs of their weak learners, usually VFDTs. To reduce the memory cost, keeping the predictive performance, this study proposes the Strict VFDT (SVFDT), a novel algorithm based on the VFDT. The SVFDT algorithm minimises unnecessary tree growth, substantially reducing memory usage and keeping competitive predictive performance. Moreover, since it creates much more shallow trees than VFDT, SVFDT can achieve a shorter processing time. Experiments were carried out comparing the SVFDT with the VFDT in 11 benchmark data stream datasets. This comparison assessed the trade-off between accuracy, memory, and processing time. Statistical analysis showed that the proposed algorithm obtained similar predictive performance and significantly reduced processing time and memory use. Thus, SVFDT is a suitable option for data stream mining with memory and time limitations, recommended as a weak learner in ensemble-based solutions.
△ Less
Submitted 17 May, 2018; v1 submitted 16 May, 2018;
originally announced May 2018.
-
Disentangling magnetic hardening and molecular spin chain contributions to exchange bias in ferromagnet/molecule bilayers
Authors:
Samy Boukari,
Hashim Jabbar,
Filip Schleicher,
Manuel Gruber,
Jacek Arabski,
Victor Da Costa,
Guy Schmerber,
Prashanth Rengasamy,
Bertrand Vileno,
Wolfgang Weber,
Martin Bowen,
Eric Beaurepaire
Abstract:
We performed SQUID and FMR magnetometry experiments to clarify the relationship between two reported magnetic exchange effects arising from interfacial spin-polarized charge transfer within ferromagnetic metal (FM)/molecule bilayers: the magnetic hardening effect, and spinterface-stabilized molecular spin chains. To disentangle these effects, both of which can affect the FM magnetization reversal,…
▽ More
We performed SQUID and FMR magnetometry experiments to clarify the relationship between two reported magnetic exchange effects arising from interfacial spin-polarized charge transfer within ferromagnetic metal (FM)/molecule bilayers: the magnetic hardening effect, and spinterface-stabilized molecular spin chains. To disentangle these effects, both of which can affect the FM magnetization reversal, we tuned the metal phthalocyanine molecule central site's magnetic moment to selectively enhance or suppress the formation of spin chains within the molecular film. We find that both effects are distinct, and additive. In the process, we 1) extended the list of FM/molecule candidate pairs that are known to generate magnetic exchange effects, 2) experimentally confirmed the predicted increase in anisotropy upon molecular adsorption; and 3) showed that spin chains within the molecular film can enhance magnetic exchange. This magnetic ordering within the organic layer implies a structural ordering. Thus, by distengangling the magnetic hardening and exchange bias contributions, our results confirm, as an echo to progress regarding inorganic spintronic tunnelling, that the milestone of spintronic tunnelling across structurally ordered organic barriers has been reached through previous magnetotransport experiments. This paves the way for solid-state devices studies that exploit the quantum physical properties of spin chains, notably through external stimuli.
△ Less
Submitted 20 December, 2017;
originally announced December 2017.
-
Light controlled magnetoresistance and magnetic field controlled photoresistance in CoFe film deposited on BiFeO3
Authors:
B. Kundys,
C. Meny,
M. R. J. Gibbs,
V. Da Costa,
M. Viret,
M. Acosta,
D. Colson,
B. Doudin
Abstract:
We present a magnetoresistive-photoresistive device based on the interaction of a piezomagnetic CoFe thin film with a photostrictive BiFeO3 substrate that undergoes light-induced strain. The magnitude of the resistance and magnetoresistance in the CoFe film can be controlled by the wavelength of the incident light on the BiFeO3. Moreover, a light-induced decrease in anisotropic magnetoresistance i…
▽ More
We present a magnetoresistive-photoresistive device based on the interaction of a piezomagnetic CoFe thin film with a photostrictive BiFeO3 substrate that undergoes light-induced strain. The magnitude of the resistance and magnetoresistance in the CoFe film can be controlled by the wavelength of the incident light on the BiFeO3. Moreover, a light-induced decrease in anisotropic magnetoresistance is detected due to an additional magnetoelastic contribution to magnetic anisotropy of the CoFe film. This effect may find applications in photo-sensing systems, wavelength detectors and can possibly open a research development in light-controlled magnetic switching properties for next generation magnetoresistive memory devices.
△ Less
Submitted 2 July, 2012;
originally announced July 2012.
-
Photostriction in BiFeO3: wavelength dependence
Authors:
B. Kundys,
M. Viret,
C. Meny,
V. Da Costa,
D. Colson,
B. Doudin
Abstract:
In electrically polar solids optomechanical effects result from the combination of two main processes, electric field-induced strain and photon-induced voltages. Whereas the former depends on the electrostrictive ability of the sample to convert electric energy into mechanical energy, the latter is caused by the capacity of photons with appropriate energy to generate charges and, therefore, can de…
▽ More
In electrically polar solids optomechanical effects result from the combination of two main processes, electric field-induced strain and photon-induced voltages. Whereas the former depends on the electrostrictive ability of the sample to convert electric energy into mechanical energy, the latter is caused by the capacity of photons with appropriate energy to generate charges and, therefore, can depend on wavelength.We report here on mechanical deformation of BiFeO3 and its response time to discrete wavelengths of incident light ranging from 365 to 940 nm. The mechanical response of BiFeO3 is found to have two maxima in near-UV and green spectral wavelength regions.
△ Less
Submitted 28 March, 2012;
originally announced March 2012.
-
Broad distribution effects in sums of lognormal random variables
Authors:
M. Romeo,
V. Da Costa,
F. Bardou
Abstract:
The lognormal distribution describing, e.g., exponentials of Gaussian random variables is one of the most common statistical distributions in physics. It can exhibit features of broad distributions that imply qualitative departure from the usual statistical scaling associated to narrow distributions. Approximate formulae are derived for the typical sums of lognormal random variables. The validit…
▽ More
The lognormal distribution describing, e.g., exponentials of Gaussian random variables is one of the most common statistical distributions in physics. It can exhibit features of broad distributions that imply qualitative departure from the usual statistical scaling associated to narrow distributions. Approximate formulae are derived for the typical sums of lognormal random variables. The validity of these formulae is numerically checked and the physical consequences, e.g., for the current flowing through small tunnel junctions, are pointed out.
△ Less
Submitted 18 March, 2003; v1 submitted 14 November, 2002;
originally announced November 2002.
-
Statistical properties of currents flowing through tunnel junctions
Authors:
V. Da Costa,
M. Romeo,
F. Bardou
Abstract:
This paper presents an overview of the statistical properties arising from the broadness of the distribution of tunnel currents in metal-insulator-metal tunnel junctions. Experimental current inhomogeneities can be modelled by a lognormal distribution and the size dependence of the tunnel current is modified at small sizes by the effect of broad distributions.
This paper presents an overview of the statistical properties arising from the broadness of the distribution of tunnel currents in metal-insulator-metal tunnel junctions. Experimental current inhomogeneities can be modelled by a lognormal distribution and the size dependence of the tunnel current is modified at small sizes by the effect of broad distributions.
△ Less
Submitted 24 July, 2002; v1 submitted 23 May, 2002;
originally announced May 2002.
