arXiv subjects

Eduardo Ros

Publications and source records attributed to Eduardo Ros.

At least 55 records · Page 3Linked to original sources

arXiv2020

Simulation, visualization and analysis tools for pattern recognition assessment with spiking neuronal networks

Computational modeling is becoming a widely used methodology in modern neuroscience. However, as the complexity of the phenomena under study increases, the analysis of the results emerging from the simulations concomitantly becomes more challenging. In particular, the configuration and validation of brain circuits involving learning often require the processing of large amounts of action potentials and their comparison to the stimulation being presented to the input of the system. In this study we present a systematic work-flow for the configuration of spiking-neuronal-network-based learning systems including evolutionary algorithms for information transmission optimization, advanced visualization tools for the validation of the best suitable configuration and customized scripts for final quantitative evaluation of the learning capabilities. By integrating both grouped action potential information and stimulation-related events, the proposed visualization framework provides qualitatively assessment of the evolution of the learning process in the simulation under study. The proposed work-flow has been used to study how receptive fields emerge in a network of inhibitory interneurons with excitatory and inhibitory spike-timing dependent plasticity when it is exposed to repetitive and partially overlapped stimulation patterns. According to our results, the output population reliably detected the presence of the stimulation patterns, even when the fan-in ratio of the interneurons was considerably restricted.

q-bio.NC↗

arXiv2020

Exploring vestibulo-ocular adaptation in a closed-loop neuro-robotic experiment using STDP. A simulation study

Studying and understanding the computational primitives of our neural system requires for a diverse and complementary set of techniques. In this work, we use the Neuro-robotic Platform (NRP)to evaluate the vestibulo ocular cerebellar adaptatIon (Vestibulo-ocular reflex, VOR)mediated by two STDP mechanisms located at the cerebellar molecular layer and the vestibular nuclei respectively. This simulation study adopts an experimental setup (rotatory VOR)widely used by neuroscientists to better understand the contribution of certain specific cerebellar properties (i.e. distributed STDP, neural properties, coding cerebellar topology, etc.)to r-VOR adaptation. The work proposes and describes an embodiment solution for which we endow a simulated humanoid robot (iCub)with a spiking cerebellar model by means of the NRP, and we face the humanoid to an r-VOR task. The results validate the adaptive capabilities of the spiking cerebellar model (with STDP)in a perception-action closed-loop (r- VOR)causing the simulated iCub robot to mimic a human behavior.

q-bio.NC↗

arXiv2020

A Metric for Evaluating Neural Input Representation in Supervised Learning Networks

Supervised learning has long been attributed to several feed-forward neural circuits within the brain, with attention being paid to the cerebellar granular layer. The focus of this study is to evaluate the input activity representation of these feed-forward neural networks. The activity of cerebellar granule cells is conveyed by parallel fibers and translated into Purkinje cell activity; the sole output of the cerebellar cortex. The learning process at this parallel-fiber-to-Purkinje-cell connection makes each Purkinje cell sensitive to a set of specific cerebellar states, determined by the granule-cell activity during a certain time window. A Purkinje cell becomes sensitive to each neural input state and, consequently, the network operates as a function able to generate a desired output for each provided input by means of supervised learning. However, not all sets of Purkinje cell responses can be assigned to any set of input states due to the network's own limitations (inherent to the network neurobiological substrate), that is, not all input-output mapping can be learned. A limiting factor is the representation of the input states through granule-cell activity. The quality of this representation will determine the capacity of the network to learn a varied set of outputs. In this study we present an algorithm for evaluating quantitatively the level of compatibility/interference amongst a set of given cerebellar states according to their representation (granule-cell activation patterns) without the need for actually conducting simulations and network training. The algorithm input consists of a real-number matrix that codifies the activity level of every considered granule-cell in each state. The capability of this representation to generate a varied set of outputs is evaluated geometrically, thus resulting in a real number that assesses the goodness of the representation

cs.NE↗

arXiv2019

Studying Black Holes on Horizon Scales with VLBI Ground Arrays

High-resolution imaging of supermassive black holes is now possible, with new applications to testing general relativity and horizon-scale accretion and relativistic jet formation processes. Over the coming decade, the EHT will propose to add new strategically placed VLBI elements operating at 1.3mm and 0.87mm wavelength. In parallel, development of next-generation backend instrumentation, coupled with high throughput correlation architectures, will boost sensitivity, allowing the new stations to be of modest collecting area while still improving imaging fidelity and angular resolution. The goal of these efforts is to move from imaging static horizon scale structure to dynamic reconstructions that capture the processes of accretion and jet launching in near real time.

astro-ph.IM↗

arXiv2019

First M87 Event Horizon Telescope Results and the Role of ALMA

In April 2019, the Event Horizon Telescope (EHT) collaboration revealed the first image of the candidate super-massive black hole (SMBH) at the centre of the giant elliptical galaxy Messier 87 (M87). This event-horizon-scale image shows a ring of glowing plasma with a dark patch at the centre, which is interpreted as the shadow of the black hole. This breakthrough result, which represents a powerful confirmation of Einstein's theory of gravity, or general relativity, was made possible by assembling a global network of radio telescopes operating at millimetre wavelengths that for the first time included the Atacama Large Millimeter/ submillimeter Array (ALMA). The addition of ALMA as an anchor station has enabled a giant leap forward by increasing the sensitivity limits of the EHT by an order of magnitude, effectively turning it into an imaging array. The published image demonstrates that it is now possible to directly study the event horizon shadows of SMBHs via electromagnetic radiation, thereby transforming this elusive frontier from a mathematical concept into an astrophysical reality. The expansion of the array over the next few years will include new stations on different continents - and eventually satellites in space. This will provide progressively sharper and higher-fidelity images of SMBH candidates, and potentially even movies of the hot plasma orbiting around SMBHs. These improvements will shed light on the processes of black hole accretion and jet formation on event-horizon scales, thereby enabling more precise tests of general relativity in the truly strong field regime.

astro-ph.HE↗

arXiv2018

Global millimeter VLBI array survey of ultracompact extragalactic radio sources at 86 GHz

(abridged) Very long baseline interferometry (VLBI) observations at 86$\,$GHz (wavelength, $λ= 3\,$mm) reach a resolution of about 50 $μ$as, probing the collimation and acceleration regions of relativistic outflows in active galactic nuclei. To extend the statistical studies of compact extragalactic jets, a large global 86 GHz VLBI survey of 162 radio sources was conducted in 2010-2011 using the Global Millimeter VLBI Array. The survey data attained a typical baseline sensitivity of 0.1 Jy and a typical image sensitivity of 5 mJy/beam, providing successful detections and images for all of the survey targets. For 138 objects, the survey provides the first ever VLBI images made at 86 GHz. Gaussian model fitting of the visibility data was applied to represent the structure of the sources. The Gaussian model-fit-based estimates of brightness temperature ($T_\mathrm{b}$) at the jet base (core) and in moving regions (jet components) downstream from the core were compared to the estimates of $T_\mathrm{b}$ limits made directly from the visibility data, demonstrating a good agreement between the two methods. The apparent brightness temperature estimates for the jet cores in our sample range from $2.5 \times 10^{9}\,$K to $ 1.3\times 10^{12}\,$K. A population model with a single intrinsic value of brightness temperature, $T_\mathrm{0}$, is applied to reproduce the observed $T_\mathrm{b}$ distribution. It yields $T_\mathrm{0} = (3.77^{+0.10}_{-0.14}) \times 10^{11}\,$K for the jet cores, implying that the inverse Compton losses dominate the emission. In the jet components, $T_\mathrm{0} =(1.42^{+0.16}_{-0.19})\times 10^{11}\,$K is found, slightly higher than the equipartition limit of $\sim5\times 10^{10}\,$K expected for these jet regions. For objects with sufficient structural detail detected, the adiabatic energy losses dominate the observed changes of $T_\mathrm{b}$ along the jet.

astro-ph.GA↗

arXiv2018

Real-time clustering and multi-target tracking using event-based sensors

Clustering is crucial for many computer vision applications such as robust tracking, object detection and segmentation. This work presents a real-time clustering technique that takes advantage of the unique properties of event-based vision sensors. Since event-based sensors trigger events only when the intensity changes, the data is sparse, with low redundancy. Thus, our approach redefines the well-known mean-shift clustering method using asynchronous events instead of conventional frames. The potential of our approach is demonstrated in a multi-target tracking application using Kalman filters to smooth the trajectories. We evaluated our method on an existing dataset with patterns of different shapes and speeds, and a new dataset that we collected. The sensor was attached to the Baxter robot in an eye-in-hand setup monitoring real-world objects in an action manipulation task. Clustering accuracy achieved an F-measure of 0.95, reducing the computational cost by 88% compared to the frame-based method. The average error for tracking was 2.5 pixels and the clustering achieved a consistent number of clusters along time.

cs.RO↗

arXiv2018

Detection of intrinsic source structure at ~3 Schwarzschild radii with Millimeter-VLBI observations of SAGITTARIUS A*

We report results from very long baseline interferometric (VLBI) observations of the supermassive black hole in the Galactic center, Sgr A*, at 1.3 mm (230 GHz). The observations were performed in 2013 March using six VLBI stations in Hawaii, California, Arizona, and Chile. Compared to earlier observations, the addition of the APEX telescope in Chile almost doubles the longest baseline length in the array, provides additional {\it uv} coverage in the N-S direction, and leads to a spatial resolution of $\sim$30 $μ$as ($\sim$3 Schwarzschild radii) for Sgr A*. The source is detected even at the longest baselines with visibility amplitudes of $\sim$4-13% of the total flux density. We argue that such flux densities cannot result from interstellar refractive scattering alone, but indicate the presence of compact intrinsic source structure on scales of $\sim$3 Schwarzschild radii. The measured nonzero closure phases rule out point-symmetric emission. We discuss our results in the context of simple geometric models that capture the basic characteristics and brightness distributions of disk- and jet-dominated models and show that both can reproduce the observed data. Common to these models are the brightness asymmetry, the orientation, and characteristic sizes, which are comparable to the expected size of the black hole shadow. Future 1.3 mm VLBI observations with an expanded array and better sensitivity will allow a more detailed imaging of the horizon-scale structure and bear the potential for a deep insight into the physical processes at the black hole boundary.

astro-ph.GA↗

arXiv2018

The 1.4 mm core of Centaurus A: First VLBI results with the South Pole Telescope

Centaurus A (Cen A) is a bright radio source associated with the nearby galaxy NGC 5128 where high-resolution radio observations can probe the jet at scales of less than a light-day. The South Pole Telescope (SPT) and the Atacama Pathfinder Experiment (APEX) performed a single-baseline very-long-baseline interferometry (VLBI) observation of Cen A in January 2015 as part of VLBI receiver deployment for the SPT. We measure the correlated flux density of Cen A at a wavelength of 1.4 mm on a $\sim$7000 km (5 G$λ$) baseline. Ascribing this correlated flux density to the core, and with the use of a contemporaneous short-baseline flux density from a Submillimeter Array observation, we infer a core brightness temperature of $1.4 \times 10^{11}$ K. This is close to the equipartition brightness temperature, where the magnetic and relativistic particle energy densities are equal. Under the assumption of a circular Gaussian core component, we derive an upper limit to the core size $ϕ= 34.0 \pm 1.8~μ\textrm{as}$, corresponding to 120 Schwarzschild radii for a black hole mass of $5.5 \times 10^7 M_{\odot}$.

astro-ph.HE↗

arXiv2018

Joint direct estimation of 3D geometry and 3D motion using spatio temporal gradients

Conventional image motion based structure from motion methods first compute optical flow, then solve for the 3D motion parameters based on the epipolar constraint, and finally recover the 3D geometry of the scene. However, errors in optical flow due to regularization can lead to large errors in 3D motion and structure. This paper investigates whether performance and consistency can be improved by avoiding optical flow estimation in the early stages of the structure from motion pipeline, and it proposes a new direct method based on image gradients (normal flow) only. The main idea lies in a reformulation of the positive-depth constraint, which allows the use of well-known minimization techniques to solve for 3D motion. The 3D motion estimate is then refined and structure estimated adding a regularization based on depth. Experimental comparisons on standard synthetic datasets and the real-world driving benchmark dataset KITTI using three different optic flow algorithms show that the method achieves better accuracy in all but one case. Furthermore, it outperforms existing normal flow based 3D motion estimation techniques. Finally, the recovered 3D geometry is shown to be also very accurate.

cs.CV↗

arXiv2018

Community Paper on Radio Astronomy Infrastructures

Radio astronomy has experienced phenomenal progress in recent years due to advances in digital technologies and processing speed, the development of new technologies, and the prospect for new powerful facilities A new generation of radio interferometers is opening new windows to the Universe. LOFAR, the world largest telescope, extended the classical radio window at low frequencies. ALMA started a new era at millimetre wavelengths and the planned SKA will revolutionize the sciences of the Universe, well beyond the traditional limits of astronomy. Radio astronomy research is making leaps with enhancements in resolution, sensitivity, and image fidelity. In Germany, the community has access to excellent facilities and training opportunities. The Effelsberg telescope remains the flagship of radio astronomical research at centimeter wavelengths and serves as a test bed for new technologies, LOFAR has notably expanded the German community and now includes six German stations and a LOFAR long term archive coordinated by the GLOW consortium. The SKA will be a transformational astronomical facility in the coming decade(s), and the German community is looking forward to broadly participate in SKA-enabled research.

astro-ph.IM↗

arXiv2016

VLBI studies of TANAMI radio galaxies

Radio galaxies are relatively faint at $γ$-ray energies, where they make up only 1-2% of all AGN detected by Fermi-LAT. However, they offer a unique perspective to study the intrinsic properties of AGN jets. For this reason, the combination of $γ$-ray and multi-wavelength data with high-resolution VLBI monitoring is a powerful tool to tackle the basic unanswered questions about AGN jets. Here we present preliminary results from a sample study of radio galaxies in the Southern hemisphere observed by the TANAMI VLBI monitoring program. We obtain high-resolution maps at 8.4 and 22.3 GHz, and study the jet kinematics using multi-epoch data. We present a preliminary kinematic analysis for the peculiar $γ$-ray AGN PKS 0521$-$36.

astro-ph.HE↗

arXiv2016

The TANAMI Multiwavelength Program: Dynamic SEDs of Southern Blazars

Simultaneous broadband spectral and temporal studies of blazars are an important tool for investigating active galactic nuclei (AGN) jet physics. We study the spectral evolution between quiescent and flaring periods of 22 radio-loud AGN through multi-epoch, quasi-simultaneous broadband spectra. For many of these sources these are the first broadband studies. We use a Bayesian block analysis of \Fermi/LAT light curves in order to determine time ranges of constant flux for constructing quasi-simultaneous SEDs. The shapes of the resulting 81 SEDs are described by two logarithmic parabolas and a blackbody spectrum where needed. For low states the peak frequencies and luminosities agree well with the blazar sequence, higher luminosity implying lower peak frequencies. This is not true for sources in a high state. The $γ$-ray photon index in Fermi/LAT correlates with the synchrotron peak frequency in low and intermediate states. No correlation is present in high states. The black hole mass cannot be determined from the SEDs. Surprisingly, the thermal excess often found in FSRQs at optical/UV wavelengths can be described by blackbody emission and not an accretion disk spectrum. The "harder-when-brighter" trend, typically seen in X-ray spectra of flaring blazars, is visible in the blazar sequence. Our results for low and intermediate states, as well as the Compton dominance, are in agreement with previous results. Black hole mass estimates using the parameters from Bonchi (2013) are in agreement with some of the more direct measurements. For two sources, estimates disagree by more than four orders of magnitude, possibly due to boosting effects. The shapes of the thermal excess seen predominantly in flat spectrum radio quasars are inconsistent with a direct accretion disk origin.

astro-ph.HE↗

arXiv2016

Scalability in Neural Control of Musculoskeletal Robots

Anthropomimetic robots are robots that sense, behave, interact and feel like humans. By this definition, anthropomimetic robots require human-like physical hardware and actuation, but also brain-like control and sensing. The most self-evident realization to meet those requirements would be a human-like musculoskeletal robot with a brain-like neural controller. While both musculoskeletal robotic hardware and neural control software have existed for decades, a scalable approach that could be used to build and control an anthropomimetic human-scale robot has not been demonstrated yet. Combining Myorobotics, a framework for musculoskeletal robot development, with SpiNNaker, a neuromorphic computing platform, we present the proof-of-principle of a system that can scale to dozens of neurally-controlled, physically compliant joints. At its core, it implements a closed-loop cerebellar model which provides real-time low-level neural control at minimal power consumption and maximal extensibility: higher-order (e.g., cortical) neural networks and neuromorphic sensors like silicon-retinae or -cochleae can naturally be incorporated.

cs.RO↗

arXiv2015

The Gamma-Ray Emitting Radio-Loud Narrow-Line Seyfert 1 Galaxy PKS 2004-447 I. The X-Ray View

As part of the TANAMI multiwavelength progam, we discuss new X-ray observations of the $γ$-ray and radio-loud Narrow Line Seyfert galaxy ($γ$-NLS1) PKS 2004-447. The active galaxy is a member of a small sample of radio-loud NLS1s detected in $γ$-rays by the Fermi Large Area Telescope. It is the radio-loudest and only southern-hemisphere source in this sample. We present results from our X-ray monitoring program comprised of Swift snapshot observations from 2012 through 2014 and two new X-ray observations with XMM-Newton in 2012. We analyze the X-ray spectrum and variability of this peculiar source using supplementary archival data from 2004 and 2011. The (0.5-10) keV spectrum is well described by a flat power law, which can be interpreted as non-thermal emission from a relativistic jet. The source exhibits moderate flux variability on timescales of both months and years. Correlated brightness variations in the (0.5-2) keV and (2-10) keV bands are explained by a single variable spectral component, such as the jet. A possible soft excess seen in the data from 2004 cannot be confirmed by the new \xmm{} observations in 2012. Any contribution to the total flux in 2004 is less than $20\%$ of the power-law component. The (0.5-10) keV luminosities of PKS 2004-447 are in the range of (0.5--2.7)$\times10^{44}\,\mathrm{erg\,s}^{-1}$. A comparison of the X-ray properties among the known $γ$-NLS1 galaxies shows that X-ray spectrum is typically dominated by a flat power law without intrinsic absorption. These objects are moderately variable in their brightness, while spectral variability is observed in at least two sources. The major difference across the X-ray spectra of $γ$-NLS1s is the luminosity, which spans a range of almost two orders of magnitude, from $10^{44}\,\mathrm{erg\,s}^{-1}$ to $10^{46}\,\mathrm{erg\,s}^{-1}$ in the (0.5-10) keV band.

astro-ph.HE↗

arXiv2015

Black Hole Lightning from the Peculiar Gamma-Ray Loud Active Galactic Nucleus IC 310

The nearby active galaxy IC 310, located in the outskirts of the Perseus cluster of galaxies is a bright and variable multi-wavelength emitter from the radio regime up to very high gamma-ray energies above 100 GeV. Originally, the nucleus of IC 310 has been classified as a radio galaxy. However, studies of the multi-wavelength emission showed several properties similarly to those found from blazars as well as radio galaxies. In late 2012, we have organized the first contemporaneous multi-wavelength campaign including radio, optical, X-ray and gamma-ray instruments. During this campaign an exceptionally bright flare of IC 310 was detected with the MAGIC telescopes in November 2012 reaching an averaged flux level in the night of up to one Crab above 1 TeV with a hard spectrum over two decades in energy. The intra-night light curve showed a series of strong outbursts with flux-doubling time scales as fast as a few minutes. The fast variability constrains the size of the gamma-ray emission regime to be smaller than 20% of the gravitational radius of its central black hole. This challenges the shock acceleration models, commonly used to explain gamma-ray radiation from active galaxies. Here, we will present more details on the MAGIC data and discuss several possible alternative emission models.

astro-ph.HE↗

arXiv2015

The off-axis jet structure in Mrk 501 at mm-wavelengths

We present results from 43 GHz (VLBA, six epochs from 2012.2 to 2013.2) and 86 GHz (GMVA, one epoch in 2012.4) observations toward the basis of the jet in the TeV Blazar Mrk 501. The 43-GHz data analysis reveals a new feature located northeast of the radio core, with a flux density of several tens of mJy, perpendicularly to the jet axis. The 86-GHz image shows the jet feature located 0.75 mas southeast of the radio core, which is consistent with the previous result. The location of Gaussian model for 0.75 mas feature does not coincide with those for the jets in the 43-GHz image, however, a distribution of emission is found. We also discuss the spectral indices of the core, the northeast feature, and the jet feature between 43 GHz and 86 GHz, which show flat-to-steep, steep, and flat-to-invert, respectively.

astro-ph.HE↗

arXiv2015

Study of single top production at high energy electron positron colliders

The effect of single top production on the study of top quark pair production in future high energy electron--positron colliders is evaluated. The rate of the single top quark production process is sizeable throughout a large range of center-of-mass energies and cannot easily be distinguished from the dominant pair production process. We discuss the impact on the top quark mass extraction from a scan through the pair production threshold and the determination of top quark form factors in the continuum. These results advocate for the exploration of the inclusive $e^+e^-\rightarrow W^+bW^-\bar{b}$ process, that includes both top quark pair and single top quark production.

hep-ex↗