Showing 1–50 of 143 results for author: Subramanian, K

A Super-human Vision-based Reinforcement Learning Agent for Autonomous Racing in Gran Turismo

Authors: Miguel Vasco, Takuma Seno, Kenta Kawamoto, Kaushik Subramanian, Peter R. Wurman, Peter Stone

Abstract: Racing autonomous cars faster than the best human drivers has been a longstanding grand challenge for the fields of Artificial Intelligence and robotics. Recently, an end-to-end deep reinforcement learning agent met this challenge in a high-fidelity racing simulator, Gran Turismo. However, this agent relied on global features that require instrumentation external to the car. This paper introduces,… ▽ More Racing autonomous cars faster than the best human drivers has been a longstanding grand challenge for the fields of Artificial Intelligence and robotics. Recently, an end-to-end deep reinforcement learning agent met this challenge in a high-fidelity racing simulator, Gran Turismo. However, this agent relied on global features that require instrumentation external to the car. This paper introduces, to the best of our knowledge, the first super-human car racing agent whose sensor input is purely local to the car, namely pixels from an ego-centric camera view and quantities that can be sensed from on-board the car, such as the car's velocity. By leveraging global features only at training time, the learned agent is able to outperform the best human drivers in time trial (one car on the track at a time) races using only local input features. The resulting agent is evaluated in Gran Turismo 7 on multiple tracks and cars. Detailed ablation experiments demonstrate the agent's strong reliance on visual inputs, making it the first vision-based super-human car racing agent. △ Less

Submitted 18 June, 2024; originally announced June 2024.

Comments: Accepted at the Reinforcement Learning Conference (RLC) 2024

Using 3-D LiDAR Data for Safe Physical Human-Robot Interaction

Authors: Sarthak Arora, Karthik Subramanian, Odysseus Adamides, Ferat Sahin

Abstract: This paper explores the use of 3D lidar in a physical Human-Robot Interaction (pHRI) scenario. To achieve the aforementioned, experiments were conducted to mimic a modern shop-floor environment. Data was collected from a pool of seventeen participants while performing pre-determined tasks in a shared workspace with the robot. To demonstrate an end-to-end case; a perception pipeline was developed t… ▽ More This paper explores the use of 3D lidar in a physical Human-Robot Interaction (pHRI) scenario. To achieve the aforementioned, experiments were conducted to mimic a modern shop-floor environment. Data was collected from a pool of seventeen participants while performing pre-determined tasks in a shared workspace with the robot. To demonstrate an end-to-end case; a perception pipeline was developed that leverages reflectivity, signal, near-infrared, and point-cloud data from a 3-D lidar. This data is then used to perform safety based control whilst satisfying the speed and separation monitoring (SSM) criteria. In order to support the perception pipeline, a state-of-the-art object detection network was leveraged and fine-tuned by transfer learning. An analysis is provided along with results of the perception and the safety based controller. Additionally, this system is compared with the previous work. △ Less

Submitted 2 June, 2024; originally announced June 2024.

Comments: Submitted to IEEE-CASE 2024. Under Review

Understanding the radio luminosity function of star-forming galaxies and its cosmological evolution

Authors: Charles Jose, Luke Chamandy, Anvar Shukurov, Kandaswamy Subramanian, Luiz Felippe S. Rodrigues, Carlton M. Baugh

Abstract: We explore the redshift evolution of the radio luminosity function (RLF) of star-forming galaxies using GALFORM, a semi-analytic model of galaxy formation and a dynamo model of the magnetic field evolving in a galaxy. Assuming energy equipartition between the magnetic field and cosmic rays, we derive the synchrotron luminosity of each sample galaxy. In a model where the turbulent speed is correlat… ▽ More We explore the redshift evolution of the radio luminosity function (RLF) of star-forming galaxies using GALFORM, a semi-analytic model of galaxy formation and a dynamo model of the magnetic field evolving in a galaxy. Assuming energy equipartition between the magnetic field and cosmic rays, we derive the synchrotron luminosity of each sample galaxy. In a model where the turbulent speed is correlated with the star formation rate, the RLF is in fair agreement with observations in the redshift range $0 \leq z \leq 2$. At larger redshifts, the structure of galaxies, their interstellar matter and turbulence appear to be rather different from those at $z\lesssim2$, so that the turbulence and magnetic field models applicable at low redshifts become inadequate. The strong redshift evolution of the RLF at $0 \leq z \leq 2$ can be attributed to an increased number, at high redshift, of galaxies with large disc volumes and strong magnetic fields. On the other hand, in models where the turbulent speed is a constant or an explicit function of $z$, the observed redshift evolution of the RLF is poorly captured. The evolution of the interstellar turbulence and outflow parameters appear to be major (but not the only) drivers of the RLF changes. We find that both the small- and large-scale magnetic fields contribute to the RLF but the small-scale field dominates at high redshifts. Polarisation observations will therefore be important to distinguish these two components and understand better the evolution of galaxies and their nonthermal constituents. △ Less

Submitted 11 June, 2024; v1 submitted 23 February, 2024; originally announced February 2024.

Comments: 19 pages, 8 figures. MNRAS, in press

Extragalactic Magnetism with SOFIA (SALSA Legacy Program). VII. A Tomographic View of Far-infrared and Radio Polarimetric Observations through MHD Simulations of Galaxies

Authors: Sergio Martin-Alvarez, Enrique Lopez-Rodriguez, Tara Dacunha, Susan E. Clark, Alejandro S. Borlaff, Rainer Beck, Francisco Rodríguez Montero, S. Lyla Jung, Julien Devriendt, Adrianne Slyz, Julia Roman-Duval, Evangelia Ntormousi, Mehrnoosh Tahani, Kandaswamy Subramanian, Daniel A. Dale, Pamela M. Marcum, Konstantinos Tassis, Ignacio del Moral-Castro, Le Ngoc Tram, Matt J. Jarvis

Abstract: The structure of magnetic fields in galaxies remains poorly constrained, despite the importance of magnetism in the evolution of galaxies. Radio synchrotron and far-infrared (FIR) polarization and polarimetric observations are the best methods to measure galactic scale properties of magnetic fields in galaxies beyond the Milky Way. We use synthetic polarimetric observations of a simulated galaxy t… ▽ More The structure of magnetic fields in galaxies remains poorly constrained, despite the importance of magnetism in the evolution of galaxies. Radio synchrotron and far-infrared (FIR) polarization and polarimetric observations are the best methods to measure galactic scale properties of magnetic fields in galaxies beyond the Milky Way. We use synthetic polarimetric observations of a simulated galaxy to identify and quantify the regions, scales, and interstellar medium (ISM) phases probed at FIR and radio wavelengths. Our studied suite of magnetohydrodynamical cosmological zoom-in simulations features high-resolutions (10 pc full-cell size) and multiple magnetization models. Our synthetic observations have a striking resemblance to those of observed galaxies. We find that the total and polarized radio emission extends to approximately double the altitude above the galactic disk (half-intensity disk thickness of $h_\text{I radio} \sim h_\text{PI radio} = 0.23 \pm 0.03$ kpc) relative to the total FIR and polarized emission that are concentrated in the disk midplane ($h_\text{I FIR} \sim h_\text{PI FIR} = 0.11 \pm 0.01$ kpc). Radio emission traces magnetic fields at scales of $\gtrsim 300$ pc, whereas FIR emission probes magnetic fields at the smallest scales of our simulations. These scales are comparable to our spatial resolution and well below the spatial resolution ($<300$ pc) of existing FIR polarimetric measurements. Finally, we confirm that synchrotron emission traces a combination of the warm neutral and cold neutral gas phases, whereas FIR emission follows the densest gas in the cold neutral phase in the simulation. These results are independent of the ISM magnetic field strength. The complementarity we measure between radio and FIR wavelengths motivates future multiwavelength polarimetric observations to advance our knowledge of extragalactic magnetism. △ Less

Submitted 24 March, 2024; v1 submitted 10 November, 2023; originally announced November 2023.

Comments: Accepted for publication in the ApJ. Replaced to match final version. 35 pages, 17 figures, 4 tables

Spatial and Temporal Attention-based emotion estimation on HRI-AVC dataset

Authors: Karthik Subramanian, Saurav Singh, Justin Namba, Jamison Heard, Christopher Kanan, Ferat Sahin

Abstract: Many attempts have been made at estimating discrete emotions (calmness, anxiety, boredom, surprise, anger) and continuous emotional measures commonly used in psychology, namely `valence' (The pleasantness of the emotion being displayed) and `arousal' (The intensity of the emotion being displayed). Existing methods to estimate arousal and valence rely on learning from data sets, where an expert ann… ▽ More Many attempts have been made at estimating discrete emotions (calmness, anxiety, boredom, surprise, anger) and continuous emotional measures commonly used in psychology, namely `valence' (The pleasantness of the emotion being displayed) and `arousal' (The intensity of the emotion being displayed). Existing methods to estimate arousal and valence rely on learning from data sets, where an expert annotator labels every image frame. Access to an expert annotator is not always possible, and the annotation can also be tedious. Hence it is more practical to obtain self-reported arousal and valence values directly from the human in a real-time Human-Robot collaborative setting. Hence this paper provides an emotion data set (HRI-AVC) obtained while conducting a human-robot interaction (HRI) task. The self-reported pair of labels in this data set is associated with a set of image frames. This paper also proposes a spatial and temporal attention-based network to estimate arousal and valence from this set of image frames. The results show that an attention-based network can estimate valence and arousal on the HRI-AVC data set even when Arousal and Valence values are unavailable per frame. △ Less

Submitted 19 October, 2023; originally announced October 2023.

Measurement of isoplanatic angle and turbulence strength profile from H-alpha images of the Sun

Authors: Saraswathi Kalyani Subramanian, Sridharan Rengaswamy

Abstract: Adaptive Optics (AO) systems have become integral for ground-based astronomy. Based on the scientific case, there are various flavours of AO systems. Measuring the turbulence strength profile ($C_N^2(h)$) and other site characteristics is essential before selecting a site or implementing certain types of AO systems. We used an iterative deconvolution procedure on long-exposure H-$α$ images of the… ▽ More Adaptive Optics (AO) systems have become integral for ground-based astronomy. Based on the scientific case, there are various flavours of AO systems. Measuring the turbulence strength profile ($C_N^2(h)$) and other site characteristics is essential before selecting a site or implementing certain types of AO systems. We used an iterative deconvolution procedure on long-exposure H-$α$ images of the Sun to determine the isoplanatic patch size during the daytime. Then, we determined the relationship between turbulence along different directions and also obtained an analytical estimate of the $C_N^2(h)$ profile. △ Less

Submitted 18 October, 2023; originally announced October 2023.

Comments: Proceeding for AO4ELT7

Emergent hydrodynamic behaviour of few strongly interacting fermions

Authors: Sandra Brandstetter, Philipp Lunt, Carl Heintze, Giuliano Giacalone, Lars H. Heyen, Maciej Gałka, Keerthan Subramanian, Marvin Holten, Philipp M. Preiss, Stefan Floerchinger, Selim Jochim

Abstract: Hydrodynamics provides a successful framework to effectively describe the dynamics of complex many-body systems ranging from subnuclear to cosmological scales by introducing macroscopic quantities such as particle densities and fluid velocities. According to textbook knowledge, it requires coarse graining over microscopic constituents to define a macroscopic fluid cell, which is large compared to… ▽ More Hydrodynamics provides a successful framework to effectively describe the dynamics of complex many-body systems ranging from subnuclear to cosmological scales by introducing macroscopic quantities such as particle densities and fluid velocities. According to textbook knowledge, it requires coarse graining over microscopic constituents to define a macroscopic fluid cell, which is large compared to the interparticle spacing and the mean free path. In addition, the entire system must consist of many such fluid cells. The latter requirement on the system size has been challenged by experiments on high-energy heavy-ion collisions, where collective particle emission, typically associated with the formation of a hydrodynamic medium, has been observed with few tens of final-state particles. Here, we demonstrate emergence of hydrodynamics in a system with significantly less constituents. Our observation challenges the requirements for a hydrodynamic description, as in our system all relevant length scales, i.e. the system size, the inter-particle spacing, and the mean free path are comparable. The single particle resolution, deterministic control over particle number and interaction strength in our experiment allow us to explore the boundaries between a microscopic description and a hydrodynamic framework in unprecedented detail. △ Less

Submitted 18 August, 2023; originally announced August 2023.

Comments: 11 pages, 6 figures

Shedding light on the MRI driven dynamo in a stratified shearing box

Authors: Prasun Dhang, Abhijit Bendre, Kandaswamy Subramanian

Abstract: We study the magneto-rotational instability (MRI) driven dynamo in a geometrically thin disc ($H/R\ll 1$) using stratified zero net flux (ZNF) shearing box simulations. We find that mean fields and EMFs oscillate with a primary frequency $f_{\rm dyn} = 0.017$ ($\approx 9$ orbital period), but also have higher harmonics at $3f_{\rm dyn}$. Correspondingly, the current helicity, has two frequencies… ▽ More We study the magneto-rotational instability (MRI) driven dynamo in a geometrically thin disc ($H/R\ll 1$) using stratified zero net flux (ZNF) shearing box simulations. We find that mean fields and EMFs oscillate with a primary frequency $f_{\rm dyn} = 0.017$ ($\approx 9$ orbital period), but also have higher harmonics at $3f_{\rm dyn}$. Correspondingly, the current helicity, has two frequencies $2f_{\rm dyn}$ and $4f_{\rm dyn}$ respectively, which appear to be the beat frequencies of mean fields and EMFs as expected from the magnetic helicity density evolution equation. Further, we adopt a novel inversion algorithm called the `Iterative Removal Of Sources' (IROS), to extract the turbulent dynamo coefficients in the mean-field closure using the mean magnetic fields and EMFs obtained from the shearing box simulation. We show that an $α-$effect ($α_{yy}$) is predominantly responsible for the creation of the poloidal field from the toroidal field, while shear generates back a toroidal field from the poloidal field; indicating that an $α-Ω$-type dynamo is operative in MRI-driven accretion discs. We also find that both strong outflow ($\bar{v}_z$) and turbulent pumping ($γ_z$ ) transport mean fields away from the mid-plane. Instead of turbulent diffusivity, they are the principal sink terms in the mean magnetic energy evolution equation. We find encouraging evidence that a generative helicity flux is responsible for the effective $α$-effect. Finally, we point out potential limitations of horizontal ($x-y$) averaging in defining the `mean' on the extraction of dynamo coefficients and their physical interpretations. △ Less

Submitted 15 August, 2023; originally announced August 2023.

Comments: 16 pages, 13 figures, submitted to MNRAS, Comments are welcome

Iterative removal of sources to model the turbulent electromotive force

Authors: Abhijit B. Bendre, Jennifer Schober, Prasun Dhang, Kandaswamy Subramanian

Abstract: We describe a novel method to compute the components of dynamo tensors from direct magnetohydrodynamic (MHD) simulations. Our method relies upon an extension and generalisation of the standard Högbom CLEAN algorithm widely used in radio astronomy to systematically remove the impact of the strongest beams onto the corresponding image. This generalisation, called the Iterative Removal of Sources (IR… ▽ More We describe a novel method to compute the components of dynamo tensors from direct magnetohydrodynamic (MHD) simulations. Our method relies upon an extension and generalisation of the standard Högbom CLEAN algorithm widely used in radio astronomy to systematically remove the impact of the strongest beams onto the corresponding image. This generalisation, called the Iterative Removal of Sources (IROS) method, has been adopted here to model the turbulent electromotive force (EMF) in terms of the mean magnetic fields and currents. Analogous to the CLEAN algorithm, IROS treats the time series of the mean magnetic field and current as beams that convolve with the dynamo coefficients which are treated as (clean) images to produce the EMF time series (the dirty image). We apply this method to MHD simulations of galactic dynamos, to which we have previously employed other methods of computing dynamo coefficients such as the test-field method, the regression method, as well as local and non-local versions of the singular value decomposition (SVD) method. We show that our new method reliably recovers the dynamo coefficients from the MHD simulations. It also allows priors on the dynamo coefficients to be incorporated easily during the inversion, unlike in earlier methods. Moreover, using synthetic data, we demonstrate that it may serve as a viable post-processing tool in determining the dynamo coefficients, even when the power of additive noise to the EMF is twice as much the actual EMF. △ Less

Submitted 23 April, 2024; v1 submitted 31 July, 2023; originally announced August 2023.

Comments: 11 pages, 9 figures

Role of magnetic pressure forces in fluctuation dynamo saturation

Authors: Sharanya Sur, Kandaswamy Subramanian

Abstract: Using magnetohydrodynamic simulations of fluctuation dynamos in turbulent flows with rms Mach numbers $\mathcal{M}_{\rm rms} = 0.2, 1.1$ and $3$, we show that magnetic pressure forces play a crucial role in dynamo saturation in supersonic flows. First, as expected when pressure forces oppose compression, an increase in anticorrelation between density and magnetic field strengths obtains even in su… ▽ More Using magnetohydrodynamic simulations of fluctuation dynamos in turbulent flows with rms Mach numbers $\mathcal{M}_{\rm rms} = 0.2, 1.1$ and $3$, we show that magnetic pressure forces play a crucial role in dynamo saturation in supersonic flows. First, as expected when pressure forces oppose compression, an increase in anticorrelation between density and magnetic field strengths obtains even in subsonic flows with the anti-correlation arising from the intense but rarer magnetic structures. In supersonic flows, due to stronger compressive motions density and magnetic field strength continue to maintain a positive correlation. However, the degree of positive correlation decreases as the dynamo saturates. Secondly, we find that the unit vectors of $\nablaρ$ and $\nabla B^{2}$ are preferentially antiparallel to each other in subsonic flows. This is indicative of magnetic pressure opposing compression. This antiparallel alignment persists in transonic and supersonic flows at dynamo saturation. However, compressive motions also lead to the emergence of a parallel alignment in these flows. Finally, we consider the work done against the components of the Lorentz force and the different sources of magnetic energy growth and dissipation. We show that while in subsonic flows, suppression of field line stretching is dominant in saturating the dynamo, the picture is different in supersonic flows. Both field line stretching and compression initially amplifies the field. However, growing magnetic pressure opposes further compression of magnetic flux which tends to reduce the compressive motions. Simultaneously, field line stretching also reduces. But, suppression of compressive amplification dominates the saturation of the dynamo. △ Less

Submitted 23 November, 2023; v1 submitted 17 May, 2023; originally announced May 2023.

Comments: Accepted for publication in MNRAS

Journal ref: MNRAS, 527, 3968-3981 (2024)

Modelling Quasi-Orthographic Captures for Surface Imaging

Authors: Maniratnam Mandal, Venkatesh K. Subramanian

Abstract: Surveillance and surveying are two important applications of empirical research. A major part of terrain modelling is supported by photographic surveys which are used for capturing expansive natural surfaces using a wide range of sensors -- visual, infrared, ultrasonic, radio, etc. A natural surface is non-smooth, unpredictable and fast-varying, and it is difficult to capture all features and reco… ▽ More Surveillance and surveying are two important applications of empirical research. A major part of terrain modelling is supported by photographic surveys which are used for capturing expansive natural surfaces using a wide range of sensors -- visual, infrared, ultrasonic, radio, etc. A natural surface is non-smooth, unpredictable and fast-varying, and it is difficult to capture all features and reconstruct them accurately. An orthographic image of a surface provides a detailed holistic view capturing its relevant features. In a perfect orthographic reconstruction, images must be captured normal to each point on the surface which is practically impossible. In this paper, a detailed analysis of the constraints on imaging distance is also provided. A novel method is formulated to determine an approximate orthographic region on a surface surrounding the point of focus and additionally, some methods for approximating the orthographic boundary for faster computation is also proposed. The approximation methods have been compared in terms of computational efficiency and accuracy. △ Less

Submitted 14 May, 2023; originally announced May 2023.

Comments: arXiv admin note: substantial text overlap with arXiv:2305.08121

Extragalactic magnetism with SOFIA (SALSA Legacy Program) -- V: First results on the magnetic field orientation of galaxies

Authors: Alejandro S. Borlaff, Enrique Lopez-Rodriguez, Rainer Beck, Susan E. Clark, Evangelia Ntormousi, Konstantinos Tassis, Sergio Martin-Alvarez, Mehrnoosh Tahani, Daniel A. Dale, Ignacio del Moral Castro, Julia Roman-Duval, Pamela M. Marcum, John E. Beckman, Kandaswamy Subramanian, Sarah Eftekharzadeh, Leslie Proudfit

Abstract: We present the analysis of the magnetic field ($B$-field) structure of galaxies measured with far-infrared (FIR) and radio (3 and 6 cm) polarimetric observations. We use the first data release of the Survey on extragALactic magnetiSm with SOFIA (SALSA) of 14 nearby ($<20$ Mpc) galaxies with resolved (5 arcsec-18 arcsec; $90$ pc--$1$ kpc) imaging polarimetric observations using HAWC+/SOFIA from… ▽ More We present the analysis of the magnetic field ($B$-field) structure of galaxies measured with far-infrared (FIR) and radio (3 and 6 cm) polarimetric observations. We use the first data release of the Survey on extragALactic magnetiSm with SOFIA (SALSA) of 14 nearby ($<20$ Mpc) galaxies with resolved (5 arcsec-18 arcsec; $90$ pc--$1$ kpc) imaging polarimetric observations using HAWC+/SOFIA from $53$ to $214$ \um. We compute the magnetic pitch angle ($Ψ_{B}$) profiles as a function of the galactrocentric radius. We introduce a new magnetic alignment parameter ($ζ$) to estimate the disordered-to-ordered $B$-field ratio in spiral $B$-fields. We find FIR and radio wavelengths to not generally trace the same $B$-field morphology in galaxies. The $Ψ_{B}$ profiles tend to be more ordered with galactocentric radius in radio ($ζ_{\rm{6cm}} = 0.93\pm0.03$) than in FIR ($ζ_{\rm{154μm}} = 0.84\pm0.14$). For spiral galaxies, FIR $B$-fields are $2-75$\% more turbulent than the radio $B$-fields. For starburst galaxies, we find that FIR polarization is a better tracer of the $B$-fields along the galactic outflows than radio polarization. Our results suggest that the $B$-fields associated with dense, dusty, turbulent star-forming regions, those traced at FIR, are less ordered than warmer, less-dense regions, those traced at radio, of the interstellar medium. The FIR $B$-fields seem to be more sensitive to the activity of the star-forming regions and the morphology of the molecular clouds within a vertical height of few hundred pc in the disk of spiral galaxies than the radio $B$-fields. △ Less

Submitted 13 June, 2023; v1 submitted 23 March, 2023; originally announced March 2023.

Comments: 28 pages, 14 figures. Accepted for publication in The Astrophysical Journal

Image Quality Specification for Solar Telescopes

Authors: Saraswathi Kalyani Subramanian, Sridharan Rengaswamy

Abstract: Modern large ground-based solar telescopes are invariably equipped with adaptive optics systems to enhance the high angular resolution imaging and spectroscopic capabilities in the presence of the Earth's atmospheric turbulence. The quality of the images obtained from these telescopes can not be quantified with the Strehl ratio or other metrics that are used for nighttime astronomical telescopes d… ▽ More Modern large ground-based solar telescopes are invariably equipped with adaptive optics systems to enhance the high angular resolution imaging and spectroscopic capabilities in the presence of the Earth's atmospheric turbulence. The quality of the images obtained from these telescopes can not be quantified with the Strehl ratio or other metrics that are used for nighttime astronomical telescopes directly. In this paper, we propose to use the root mean square (rms) granulation contrast as a metric to quantify the image quality of ground-based solar telescopes. We obtain semi-logarithmic plots indicating the correspondence between the Strehl ratio and the rms granulation contrast for most practical values of the telescope diameters (D) and the atmospheric coherence diameters ($ r_0$), for various levels of adaptive optics compensation. We estimate the efficiency of a few working solar adaptive optics systems by comparing the results of our simulations with the Strehl ratio and rms granulation contrast published by these systems. Our results can be used in conjunction with a plausible 50 system efficiency to predict the lower bound on the rms granulation contrast expected from ground-based solar telescopes. △ Less

Submitted 24 January, 2023; originally announced January 2023.

Comments: 15 pages, 8 figures

Magnetic helicity fluxes from triple correlators

Authors: Kishore Gopalakrishnan, Kandaswamy Subramanian

Abstract: Fluxes of the magnetic helicity density play an important role in large-scale turbulent dynamos, allowing the growth of large-scale magnetic fields while overcoming catastrophic quenching. We show here, analytically, how several important types of magnetic helicity fluxes can arise from terms involving triple correlators of fluctuating fields in the helicity density evolution equation. For this, w… ▽ More Fluxes of the magnetic helicity density play an important role in large-scale turbulent dynamos, allowing the growth of large-scale magnetic fields while overcoming catastrophic quenching. We show here, analytically, how several important types of magnetic helicity fluxes can arise from terms involving triple correlators of fluctuating fields in the helicity density evolution equation. For this, we assume incompressibility and weak inhomogeneity, and use a quasinormal closure approximation: fourth-order correlators are replaced by products of second-order ones, and the effect of the fourth-order cumulants on the evolution of the third moments is modelled by a strong damping term. First, we show how a diffusive helicity flux, till now only measured in simulations, arises from the triple correlation term. This is accompanied by what we refer to as a `random advective flux', which predominantly transports magnetic helicity along the gradients of the random fields. We also find that a new helicity flux contribution, in some aspects similar to that first proposed by Vishniac, can arise from the triple correlator. This contribution depends on the gradients of the random magnetic and kinetic energies along the large-scale vorticity, and thus arises in any rotating, stratified system, even if the turbulence is predominantly nonhelical. It can source a large-scale dynamo by itself while spatially transporting magnetic helicity within the system. △ Less

Submitted 2 December, 2022; v1 submitted 29 September, 2022; originally announced September 2022.

Comments: 19 pages. Version accepted by ApJ. Changes from v1: More prominent discussion of prior work in sec 4.2. Full expression for the helicity fluxes (earlier equation 12) has now been moved to appendix E. Clarified the closure used. Corrected some typos in the explanation of the VC approximation (no change in the presented results)

Journal ref: ApJ 943 66 (2023)

Extragalactic magnetism with SOFIA (SALSA Legacy Program) -- IV: Program overview and first results on the polarization fraction

Authors: Enrique Lopez-Rodriguez, Sui Ann Mao, Rainer Beck, Alejandro S. Borlaff, Evangelia Ntormousi, Konstantinos Tassis, Daniel A. Dale, Julia Roman-Duval, Kandaswamy Subramanian, Sergio Martin-Alvarez, Pamela M. Marcum, Susan E. Clark, William T. Reach, Doyal A. Harper, Ellen G. Zweibel

Abstract: We present the first data release of the Survey on extragALactic magnetiSm with SOFIA (SALSA Legacy Program) with a set of 14 nearby ($<20$ Mpc) galaxies with resolved imaging polarimetric observations using HAWC+ from $53$ to $214$ $μ$m at a resolution of $5-18$" ($90$ pc $-$ $1$ kpc). We introduce the definitions and background on extragalactic magnetism, and present the scientific motivation an… ▽ More We present the first data release of the Survey on extragALactic magnetiSm with SOFIA (SALSA Legacy Program) with a set of 14 nearby ($<20$ Mpc) galaxies with resolved imaging polarimetric observations using HAWC+ from $53$ to $214$ $μ$m at a resolution of $5-18$" ($90$ pc $-$ $1$ kpc). We introduce the definitions and background on extragalactic magnetism, and present the scientific motivation and sample selection of the program. Here, we focus on the general trends in the emissive polarization fraction. Far-infrared polarimetric observations trace the thermal polarized emission of magnetically aligned dust grains across the galaxy disks with polarization fractions of $P=0-15$% in the cold, $T_{\rm d} = [19,48]$ K, and dense, $\log_{10}(N_{\rm HI+H_{2}}) = [19.96,22.91]$, interstellar medium. The spiral galaxies show a median $\langle P_{154μm} \rangle = 3.3\pm0.9 $% across the disks. We report the first polarized spectrum of starburst galaxies showing a minimum within $89-154$ $μ$m. The falling $53-154$ $μ$m polarized spectrum may be due to a decrease in the dust grain alignment efficiency produced by variations in dust temperatures along the line-of-sight in the galactic outflow. We find that the starburst galaxies and the star-forming regions within normal galaxies have the lowest polarization fractions. We find that 50% (7 out of 14) of the galaxies require a broken power-law in the $P-N_{HI+H_{2}}$ and $P-T_{d}$ relations with three different trends. Group 1 has a relative increase of anisotropic random B-fields produced by compression or shear of B-fields in the galactic outflows, starburst rings, and inner-bar of galaxies; and Groups 2 and 3 have a relative increase of isotropic random B-fields driven by star-forming regions in the spiral arms, and/or an increase of dust grain alignment efficiency caused by shock-driven regions or evolutionary stages of a galaxy. △ Less

Submitted 7 July, 2022; v1 submitted 2 May, 2022; originally announced May 2022.

Comments: 43 pages, 28 figures. Accepted for publication to ApJ

Doing More by Doing Less: How Structured Partial Backpropagation Improves Deep Learning Clusters

Authors: Adarsh Kumar, Kausik Subramanian, Shivaram Venkataraman, Aditya Akella

Abstract: Many organizations employ compute clusters equipped with accelerators such as GPUs and TPUs for training deep learning models in a distributed fashion. Training is resource-intensive, consuming significant compute, memory, and network resources. Many prior works explore how to reduce training resource footprint without impacting quality, but their focus on a subset of the bottlenecks (typically on… ▽ More Many organizations employ compute clusters equipped with accelerators such as GPUs and TPUs for training deep learning models in a distributed fashion. Training is resource-intensive, consuming significant compute, memory, and network resources. Many prior works explore how to reduce training resource footprint without impacting quality, but their focus on a subset of the bottlenecks (typically only the network) limits their ability to improve overall cluster utilization. In this work, we exploit the unique characteristics of deep learning workloads to propose Structured Partial Backpropagation(SPB), a technique that systematically controls the amount of backpropagation at individual workers in distributed training. This simultaneously reduces network bandwidth, compute utilization, and memory footprint while preserving model quality. To efficiently leverage the benefits of SPB at cluster level, we introduce JigSaw, a SPB aware scheduler, which does scheduling at the iteration level for Deep Learning Training(DLT) jobs. We find that JigSaw can improve large scale cluster efficiency by as high as 28\%. △ Less

Submitted 20 November, 2021; originally announced November 2021.

Comments: Accepted at DistributedML-2021

Observation of Cooper Pairs in a Mesoscopic 2D Fermi Gas

Authors: Marvin Holten, Luca Bayha, Keerthan Subramanian, Sandra Brandstetter, Carl Heintze, Philipp Lunt, Philipp M. Preiss, Selim Jochim

Abstract: Pairing is the fundamental requirement for fermionic superfluidity and superconductivity. To understand the mechanism behind pair formation is an ongoing challenge in the study of many strongly correlated fermionic systems. Cooper pairs are the key ingredient to BCS theory as the microscopic explanation of conventional superconductivity. They form between particles of opposite spin and momentum at… ▽ More Pairing is the fundamental requirement for fermionic superfluidity and superconductivity. To understand the mechanism behind pair formation is an ongoing challenge in the study of many strongly correlated fermionic systems. Cooper pairs are the key ingredient to BCS theory as the microscopic explanation of conventional superconductivity. They form between particles of opposite spin and momentum at the Fermi surface of the system. Here, we directly observe Cooper pairs in a mesoscopic two-dimensional Fermi gas. We apply an imaging scheme that enables us to extract the full in-situ momentum distribution of a strongly interacting Fermi gas with single particle and spin resolution. Our ultracold gas allows us to freely tune between a completely non-interacting, unpaired system and weak attractions, where we find Cooper pair correlations at the Fermi surface. When increasing the attractive interactions even further, the pairs gradually turn into deeply bound molecules breaking up the Fermi surface. Our mesoscopic system is closely related to the physics of nuclei, superconducting grains or quantum dots. With the precise control over interactions, particle number and potential landscape in our experiment, the observables we establish in this work provide a new approach to longstanding questions concerning not only such mesoscopic systems but also their connection to the macroscopic world. △ Less

Submitted 6 June, 2022; v1 submitted 23 September, 2021; originally announced September 2021.

Comments: 19 pages, 12 figures

Journal ref: Nature 606, 287-291 (2022)

Non-locality of the Turbulent Electromotive Force

Authors: Abhijit B. Bendre, Kandaswamy Subramanian

Abstract: The generation of large-scale magnetic fields ($\overline{\mathbf{B}}$) in astrophysical systems is driven by the mean turbulent electromotive force ($\overline{\cal E}$), the cross correlation between local fluctuations of velocity and magnetic fields. This can depend non-locally on $\overline{\mathbf{B}}$ through a convolution kernel $K_{ij}$. In a new approach to find $K_{ij}$, we directly fit… ▽ More The generation of large-scale magnetic fields ($\overline{\mathbf{B}}$) in astrophysical systems is driven by the mean turbulent electromotive force ($\overline{\cal E}$), the cross correlation between local fluctuations of velocity and magnetic fields. This can depend non-locally on $\overline{\mathbf{B}}$ through a convolution kernel $K_{ij}$. In a new approach to find $K_{ij}$, we directly fit the time series data of $\overline{\cal E}$ versus $\overline{\mathbf{B}}$ from a galactic dynamo simulation using singular value decomposition. We calculate the usual turbulent transport coefficients as moments of $K_{ij}$, show the importance of including non-locality over eddy length scales to fully capture their amplitudes and that higher order corrections to the standard transport coefficients are small in the present case. △ Less

Submitted 10 February, 2022; v1 submitted 22 July, 2021; originally announced July 2021.

Comments: 11 pages, 13 figures

Extragalactic magnetism with SOFIA (Legacy Program) -- II: A magnetically-driven flow in the starburst ring of NGC 1097

Authors: Enrique Lopez-Rodriguez, Rainer Beck, Susan E. Clark, Annie Hughes, Alejandro S. Borlaff, Evangelia Ntormousi, Lucas Grosset, Konstantinos Tassis, John E. Beckman, Kandaswamy Subramanian, Daniel Dale, Tanio Diaz-Santos, Legacy Team

Abstract: Galactic bars are frequent in disk galaxies and they may support the transfer of matter towards the central engine of active nuclei. The barred galaxy NGC 1097 has magnetic forces controlling the gas flow at several kpc scales, which suggest that magnetic fields (B-fields) are dynamically important along the bar and nuclear ring. However, the effect of the B-field on the gas flows in the central k… ▽ More Galactic bars are frequent in disk galaxies and they may support the transfer of matter towards the central engine of active nuclei. The barred galaxy NGC 1097 has magnetic forces controlling the gas flow at several kpc scales, which suggest that magnetic fields (B-fields) are dynamically important along the bar and nuclear ring. However, the effect of the B-field on the gas flows in the central kpc scale has not been characterized. Using thermal polarized emission at $89$ $μ$m with HAWC+/SOFIA, here, we measure that the polarized flux is spatially located at the contact regions of the outer-bar with the starburst ring. The linear polarization decomposition analysis shows that the $89$ $μ$m and radio ($3.5$ and $6.2$ cm) polarization traces two different modes, $m$, of the B-field: a constant B-field orientation and dominated by $m=0$ at $89$ $μ$m, and a spiral B-field dominated by $m=2$ at radio. We show that the B-field at 89 $μ$m is concentrated in the warmest region of a shock driven by the galactic-bar dynamics in the contact regions between the outer-bar with the starburst ring. Radio polarization traces a superposition of the spiral B-field outside and within the starburst ring. According to Faraday rotation measures between $3.5$ and $6.2$ cm, the radial component of the B-field along the contact regions points toward the galaxy's center on both sides. We conclude that gas streams outside and within the starburst ring follow the B-field, which feeds the black hole with matter from the host galaxy. △ Less

Submitted 7 October, 2021; v1 submitted 19 July, 2021; originally announced July 2021.

Comments: 17 pages, 10 figures, Accepted for publication to ApJ

Exploring Dropout Discriminator for Domain Adaptation

Authors: Vinod K Kurmi, Venkatesh K Subramanian, Vinay P. Namboodiri

Abstract: Adaptation of a classifier to new domains is one of the challenging problems in machine learning. This has been addressed using many deep and non-deep learning based methods. Among the methodologies used, that of adversarial learning is widely applied to solve many deep learning problems along with domain adaptation. These methods are based on a discriminator that ensures source and target distrib… ▽ More Adaptation of a classifier to new domains is one of the challenging problems in machine learning. This has been addressed using many deep and non-deep learning based methods. Among the methodologies used, that of adversarial learning is widely applied to solve many deep learning problems along with domain adaptation. These methods are based on a discriminator that ensures source and target distributions are close. However, here we suggest that rather than using a point estimate obtaining by a single discriminator, it would be useful if a distribution based on ensembles of discriminators could be used to bridge this gap. This could be achieved using multiple classifiers or using traditional ensemble methods. In contrast, we suggest that a Monte Carlo dropout based ensemble discriminator could suffice to obtain the distribution based discriminator. Specifically, we propose a curriculum based dropout discriminator that gradually increases the variance of the sample based distribution and the corresponding reverse gradients are used to align the source and target feature representations. An ensemble of discriminators helps the model to learn the data distribution efficiently. It also provides a better gradient estimates to train the feature extractor. The detailed results and thorough ablation analysis show that our model outperforms state-of-the-art results. △ Less

Submitted 9 July, 2021; originally announced July 2021.

Comments: This work is an extension of our BMVC-2019 paper (arXiv:1907.10628)

Extragalactic Magnetism with SOFIA (Legacy Program) -- I: The magnetic field in the multi-phase interstellar medium of M51

Authors: Alejandro S. Borlaff, Enrique Lopez-Rodriguez, Rainer Beck, Rodion Stepanov, Eva Ntormousi, Annie Hughes, Konstantinos Tassis, Pamela M. Marcum, Lucas Grosset, John E. Beckman, Leslie Proudfit, Susan E. Clark, Tanio Díaz-Santos, Sui Ann Mao, William T. Reach, Julia Roman-Duval, Kandaswamy Subramanian, Le Ngoc Tram, Ellen G. Zweibel, SOFIA Legacy Team

Abstract: The recent availability of high-resolution far-infrared (FIR) polarization observations of galaxies using HAWC+/SOFIA has facilitated studies of extragalactic magnetic fields in the cold and dense molecular disks.We investigate if any significant structural differences are detectable in the kpc-scale magnetic field of the grand design face-on spiral galaxy M51 when traced within the diffuse (radio… ▽ More The recent availability of high-resolution far-infrared (FIR) polarization observations of galaxies using HAWC+/SOFIA has facilitated studies of extragalactic magnetic fields in the cold and dense molecular disks.We investigate if any significant structural differences are detectable in the kpc-scale magnetic field of the grand design face-on spiral galaxy M51 when traced within the diffuse (radio) and the dense and cold (FIR) interstellar medium (ISM). Our analysis reveals a complex scenario where radio and FIR polarization observations do not necessarily trace the same magnetic field structure. We find that the magnetic field in the arms is wrapped tighter at 154um than at 3 and 6 cm; statistically significant lower values for the magnetic pitch angle are measured at FIR in the outskirts (R > 7 kpc) of the galaxy. This difference is not detected in the interarm region. We find strong correlations of the polarization fraction and total intensity at FIR and radio with the gas column density and 12CO(1-0) velocity dispersion. We conclude that the arms show a relative increase of small-scale turbulent B-fields at regions with increasing column density and dispersion velocities of the molecular gas. No correlations are found with HI neutral gas. The star formation rate shows a clear correlation with the radio polarized intensity, which is not found in FIR, pointing to a small-scale dynamo-driven B-field amplification scenario. This work shows that multi-wavelength polarization observations are key to disentangling the interlocked relation between star formation, magnetic fields, and gas kinematics in the multi-phase ISM. △ Less

Submitted 8 September, 2021; v1 submitted 19 May, 2021; originally announced May 2021.

Comments: 35 pages, 25 figures, 5 tables. Accepted for publication in ApJ

Small World Student Network at the University of Texas at Dallas in Times of Social Distancing

Authors: Kailash Subramanian, Joshua M. Williams, Daniel C. DeAnda, Aditya A. Agrawal, Andrei Racila, Aditi R. Prabhu, Lawrence Redlinger, Christopher Wendt, Ravi Prakash

Abstract: To limit the spread of the novel coronavirus on college campuses, a common strategy for the Fall 2020 and Spring 2021 terms has been to offer instruction weighted toward hybrid or fully online modalities. Colleges are now considering whether and how to expand hybrid or fully in-person instruction for future terms, and learn lessons from this experience for future use. Our paper uses Fall 2019 enro… ▽ More To limit the spread of the novel coronavirus on college campuses, a common strategy for the Fall 2020 and Spring 2021 terms has been to offer instruction weighted toward hybrid or fully online modalities. Colleges are now considering whether and how to expand hybrid or fully in-person instruction for future terms, and learn lessons from this experience for future use. Our paper uses Fall 2019 enrollment data for a medium-sized public American university to analyze whether some student groupings by class standing or course level are more susceptible to the spread of infectious disease through academic enrollment networks. Replicating Weeden and Cornwell [8], we find that enrollment networks at the institution are "small worlds" characterized by high clustering, short average path lengths, and multiple independent connections. Connectivity decreases as class standing (graduate vs. undergraduate; senior vs. freshman) and course level increase; as students move from generalized to specialized course loads, networks cluster by major. Holding other factors constant, policies focusing on in-person instruction for lower division students conflict with the greater risk of infectious spread through a lower division network in the absence of additional steps to minimize academic connectivity. There are academic and financial incentives for emphasizing the freshman experience, including concerns about student attrition from the first to second academic year and recouping costs of infrastructure investments in dormitories. Possible solutions could include (i) restricting face-to-face or hybrid instruction to courses in students' academic majors, which would disrupt larger networks into smaller ones and thus restrict the spread of infection across majors, and (ii) take a "scalpel" approach to instruction modes by moving online courses most likely to facilitate epidemic spread. △ Less

Submitted 2 April, 2021; originally announced April 2021.

Domain Impression: A Source Data Free Domain Adaptation Method

Authors: Vinod K Kurmi, Venkatesh K Subramanian, Vinay P Namboodiri

Abstract: Unsupervised Domain adaptation methods solve the adaptation problem for an unlabeled target set, assuming that the source dataset is available with all labels. However, the availability of actual source samples is not always possible in practical cases. It could be due to memory constraints, privacy concerns, and challenges in sharing data. This practical scenario creates a bottleneck in the domai… ▽ More Unsupervised Domain adaptation methods solve the adaptation problem for an unlabeled target set, assuming that the source dataset is available with all labels. However, the availability of actual source samples is not always possible in practical cases. It could be due to memory constraints, privacy concerns, and challenges in sharing data. This practical scenario creates a bottleneck in the domain adaptation problem. This paper addresses this challenging scenario by proposing a domain adaptation technique that does not need any source data. Instead of the source data, we are only provided with a classifier that is trained on the source data. Our proposed approach is based on a generative framework, where the trained classifier is used for generating samples from the source classes. We learn the joint distribution of data by using the energy-based modeling of the trained classifier. At the same time, a new classifier is also adapted for the target domain. We perform various ablation analysis under different experimental setups and demonstrate that the proposed approach achieves better results than the baseline models in this extremely novel scenario. △ Less

Submitted 17 February, 2021; originally announced February 2021.

Comments: Published- WACV-2021

Do Not Forget to Attend to Uncertainty while Mitigating Catastrophic Forgetting

Authors: Vinod K Kurmi, Badri N. Patro, Venkatesh K. Subramanian, Vinay P. Namboodiri

Abstract: One of the major limitations of deep learning models is that they face catastrophic forgetting in an incremental learning scenario. There have been several approaches proposed to tackle the problem of incremental learning. Most of these methods are based on knowledge distillation and do not adequately utilize the information provided by older task models, such as uncertainty estimation in predicti… ▽ More One of the major limitations of deep learning models is that they face catastrophic forgetting in an incremental learning scenario. There have been several approaches proposed to tackle the problem of incremental learning. Most of these methods are based on knowledge distillation and do not adequately utilize the information provided by older task models, such as uncertainty estimation in predictions. The predictive uncertainty provides the distributional information can be applied to mitigate catastrophic forgetting in a deep learning framework. In the proposed work, we consider a Bayesian formulation to obtain the data and model uncertainties. We also incorporate self-attention framework to address the incremental learning problem. We define distillation losses in terms of aleatoric uncertainty and self-attention. In the proposed work, we investigate different ablation analyses on these losses. Furthermore, we are able to obtain better results in terms of accuracy on standard benchmarks. △ Less

Submitted 3 February, 2021; originally announced February 2021.

Comments: Accepted WACV 2021

Journal ref: WACV 2021

Properties of polarized synchrotron emission from fluctuation-dynamo action -- I. Application to galaxy clusters

Authors: Sharanya Sur, Aritra Basu, Kandaswamy Subramanian

Abstract: Using magnetohydrodynamic simulations of fluctuation dynamos, we perform broad-bandwidth synthetic observations to investigate the properties of polarized synchrotron emission and the role that Faraday rotation plays in inferring the polarized structures in the intracluster medium (ICM) of galaxy clusters. In the saturated state of the dynamo, we find a Faraday depth (FD) dispersion… ▽ More Using magnetohydrodynamic simulations of fluctuation dynamos, we perform broad-bandwidth synthetic observations to investigate the properties of polarized synchrotron emission and the role that Faraday rotation plays in inferring the polarized structures in the intracluster medium (ICM) of galaxy clusters. In the saturated state of the dynamo, we find a Faraday depth (FD) dispersion $σ_{\rm FD} \approx 100$ rad m$^{-2}$, in agreement with observed values in the ICM. Remarkably, the FD power spectrum is qualitatively similar to $M(k)/k$, where $M(k)$ is the magnetic spectrum and $k$ the wavenumber. However, this similarity is broken at high $k$ when FD is obtained by applying RM synthesis to polarized emission from the ICM due to poor resolution and complexities of spectrum in FD space. Unlike the Gaussian probability distribution function (PDF) obtained for FD, the PDF of the synchrotron intensity is lognormal. A relatively large $σ_{\rm FD}$ in the ICM gives rise to strong frequency-dependent variations of the pixel-wise mean and peak polarized intensities at low frequencies ($\lesssim 1.5\,{\rm GHz}$). The mean fractional polarization $\langle p \rangle$ obtained at the resolution of the simulations increases from $<0.1$ at 0.5 GHz to its intrinsic value of $\sim0.3$ at 6 GHz. Beam smoothing significantly affects the polarization properties below $\lesssim 1.5\,{\rm GHz}$, reducing $\langle p \rangle$ to $\lesssim 0.01$ at 0.5 GHz. At frequencies $\gtrsim 5\,{\rm GHz}$, polarization remains largely unaffected, even when recovered using RM synthesis. Thus, our results underline the need for high-frequency ($\gtrsim 5\,{\rm GHz}$) observations with future radio telescopes to effectively probe the properties of polarized emission in the ICM. △ Less

Submitted 22 January, 2021; v1 submitted 17 November, 2020; originally announced November 2020.

Comments: Published in MNRAS

Journal ref: MNRAS, Vol. 501, Issue 3, 2021

Observation of Pauli Crystals

Authors: Marvin Holten, Luca Bayha, Keerthan Subramanian, Carl Heintze, Philipp M. Preiss, Selim Jochim

Abstract: The Pauli exclusion principle is a fundamental law underpinning the structure of matter. Due to their anti-symmetric wave function, no two fermions can occupy the same quantum state. Here, we report on the direct observation of the Pauli principle in a continuous system of up to six particles in the ground state of a two-dimensional harmonic oscillator. To this end, we sample the full many-body wa… ▽ More The Pauli exclusion principle is a fundamental law underpinning the structure of matter. Due to their anti-symmetric wave function, no two fermions can occupy the same quantum state. Here, we report on the direct observation of the Pauli principle in a continuous system of up to six particles in the ground state of a two-dimensional harmonic oscillator. To this end, we sample the full many-body wavefunction by applying a single atom resolved imaging scheme in momentum space. We find so-called Pauli crystals as a manifestation of higher order correlations. In contrast to true crystalline phases, these unique high-order density correlations emerge even without any interactions present. Our work lays the foundation for future studies of correlations in strongly interacting systems of many fermions. △ Less

Submitted 16 December, 2020; v1 submitted 8 May, 2020; originally announced May 2020.

Comments: 10 pages, 8 figures

Journal ref: Phys. Rev. Lett. 126, 020401 (2021)

Observing the emergence of a quantum phase transition -- shell by shell

Authors: Luca Bayha, Marvin Holten, Ralf Klemt, Keerthan Subramanian, Johannes Bjerlin, Stephanie M. Reimann, Georg M. Bruun, Philipp M. Preiss, Selim Jochim

Abstract: Many-body physics describes phenomena which cannot be understood looking at a systems' constituents alone. Striking manifestations are broken symmetry, phase transitions, and collective excitations. Understanding how such collective behaviour emerges when assembling a system from individual particles has been a vision in atomic, nuclear, and solid-state physics for decades. Here, we observe the fe… ▽ More Many-body physics describes phenomena which cannot be understood looking at a systems' constituents alone. Striking manifestations are broken symmetry, phase transitions, and collective excitations. Understanding how such collective behaviour emerges when assembling a system from individual particles has been a vision in atomic, nuclear, and solid-state physics for decades. Here, we observe the few-body precursor of a quantum phase transition from a normal to a superfluid phase. The transition is signalled by the softening of the mode associated with amplitude vibrations of the order parameter, commonly referred to as a Higgs mode. We achieve exquisite control over ultracold fermions confined to two-dimensional harmonic potentials and prepare closed-shell configurations of 2, 6 and 12 fermionic atoms in the ground state with high fidelity. Spectroscopy is then performed on our mesoscopic system while tuning the pair energy from zero to being larger than the shell spacing. Using full atom counting statistics, we find the lowest resonance to consist of coherently excited pairs only. The distinct non-monotonic interaction dependence of this many-body excitation as well as comparison with numerical calculations allows us to identify it as the precursor of the Higgs mode. Our atomic simulator opens new pathways to systematically unravel the emergence of collective phenomena and the thermodynamic limit particle by particle. △ Less

Submitted 30 April, 2020; originally announced April 2020.

Comments: 14 pages, 10 figures

Journal ref: Nature 587, 583-587 (2020)

Gravitational wave generation in a viable scenario of inflationary magnetogenesis

Authors: Ramkishor Sharma, Kandaswamy Subramanian, T. R. Seshadri

Abstract: Generation of magnetic fields during inflation is a promising mechanism for the origin of the observed large scale magnetic fields in the universe. Among several attempts, a popular model is one where the inflaton and the electromagnetic field are coupled through a coupling function $f$ leading to a term in the Lagrangian density of the form, $f^2 F^{μν} F_{μν}$. A number of potential difficulties… ▽ More Generation of magnetic fields during inflation is a promising mechanism for the origin of the observed large scale magnetic fields in the universe. Among several attempts, a popular model is one where the inflaton and the electromagnetic field are coupled through a coupling function $f$ leading to a term in the Lagrangian density of the form, $f^2 F^{μν} F_{μν}$. A number of potential difficulties with such models have been raised in the literature. In our earlier work, we have suggested viable models of inflationary magnetogenesis which avoid these problems and at the same time can lead to either nonhelical or helical magnetic fields of astrophysical interest. Our models require a low energy scale for inflation and reheating (reheating temperature, $T_R < 10^4$ GeV) and generate a blue spectrum of electromagnetic (EM) field which peaks around the horizon scale of reheating. We show here that the anisotropic stress associated with these EM fields naturally source the production of a stochastic background of Gravitational waves (GW) with frequencies in the range of tens of nano Hertz to milli Hertz. These two extremes of the range can be probed respectively by pulsar timing arrays (PTA) experiments and the upcoming Laser Interferometric Space Array (LISA). The peak value of the GW spectrum energy represented by $d Ω_{GW}/d \ln k$ is $ 10^{-6} $ for the models which lead to nonhelical primordial fields and $2 \times 10^{-6} $ for the helical case for $T_R=100$ GeV. In this case the spectrum peaks at a frequency $30 μ$Hz for non helical case and at $40 μ$Hz for helical case. These values are obtained when the ratio of EM energy density to the cosmological density at reheating $ε\sim 1$ and decrease approximately as $ε^2$ for smaller values. △ Less

Submitted 27 December, 2019; originally announced December 2019.

Comments: 25 pages, 5 figures, comments are welcome

Journal ref: Phys. Rev. D 101, 103526 (2020)

Characterising the Dynamo in a Radiatively Inefficient Accretion Flow

Authors: Prasun Dhang, Abhijit Bendre, Prateek Sharma, Kandaswamy Subramanian

Abstract: We explore the MRI driven dynamo in a radiatively inefficient accretion flow (RIAF) using the mean field dynamo paradigm. Using singular value decomposition (SVD) we obtain the least squares fitting dynamo coefficients $α$ and $γ$ by comparing the time series of the turbulent electromotive force and the mean magnetic field. Our study is the first one to show the poloidal distribution of these dyna… ▽ More We explore the MRI driven dynamo in a radiatively inefficient accretion flow (RIAF) using the mean field dynamo paradigm. Using singular value decomposition (SVD) we obtain the least squares fitting dynamo coefficients $α$ and $γ$ by comparing the time series of the turbulent electromotive force and the mean magnetic field. Our study is the first one to show the poloidal distribution of these dynamo coefficients in global accretion flow simulations. Surprisingly, we obtain a high value of the turbulent pumping coefficient $γ$ which transports the mean magnetic flux radially outward. This would have implications for the launching of magnetised jets which are produced efficiently in presence a large-scale poloidal magnetic field close to the compact object. We present a scenario of a truncated disc beyond the RIAF where a large scale dynamo-generated poloidal magnetic field can aid jet-launching close to the black hole. Magnitude of all the calculated coefficients decreases with radius. Meridional variations of $α_{φφ}$, responsible for toroidal to poloidal field conversion, is very similar to that found in shearing box simulations using the `test field' (TF) method. By estimating the relative importance of $α$-effect and shear, we conclude that the MRI driven large-scale dynamo, which operates at high latitudes beyond a disc scale height, is essentially of the $α-Ω$ type. △ Less

Submitted 27 April, 2020; v1 submitted 10 December, 2019; originally announced December 2019.

Comments: 13 pages, 11 figures, published in MNRAS

D2R: Dataplane-Only Policy-Compliant Routing Under Failures

Authors: Kausik Subramanian, Anubhavnidhi Abhashkumar, Loris D'Antoni, Aditya Akella

Abstract: In networks today, the data plane handles forwarding---sending a packet to the next device in the path---and the control plane handles routing---deciding the path of the packet in the network. This architecture has limitations. First, when link failures occur, the data plane has to wait for the control plane to install new routes, and packet losses can occur due to delayed routing convergence or c… ▽ More In networks today, the data plane handles forwarding---sending a packet to the next device in the path---and the control plane handles routing---deciding the path of the packet in the network. This architecture has limitations. First, when link failures occur, the data plane has to wait for the control plane to install new routes, and packet losses can occur due to delayed routing convergence or central controller latencies. Second, policy-compliance is not guaranteed without sophisticated configuration synthesis or controller intervention. In this paper, we take advantage of the recent advances in fast programmable switches to perform policy-compliant route computations entirely in the data plane, thus providing fast reactions to failures. D2R, our new network architecture, can provide the illusion of a network fabric that is always available and policy-compliant, even under failures. We implement our data plane in P4 and demonstrate its viability in real world topologies. △ Less

Submitted 5 December, 2019; originally announced December 2019.

The Origin of Large-Scale Magnetic Fields in Low-Mass Galaxies

Authors: Prasanta Bera, Anvar Shukurov, Kandaswamy Subramanian

Abstract: The origin of large-scale magnetic fields, detected in some low-mass (dwarf and irregular) galaxies via polarised synchrotron emission and Faraday rotation, remained unexplained for a long time. We suggest that mean-field dynamo can be active in galaxies of this class despite their slow rotation because their discs are relatively thick. Earlier assessments of the possibility of the mean-field dyna… ▽ More The origin of large-scale magnetic fields, detected in some low-mass (dwarf and irregular) galaxies via polarised synchrotron emission and Faraday rotation, remained unexplained for a long time. We suggest that mean-field dynamo can be active in galaxies of this class despite their slow rotation because their discs are relatively thick. Earlier assessments of the possibility of the mean-field dynamo action in low-mass galaxies relied on estimates applicable to $\textit{thin}$ discs, such as those in massive spiral galaxies. Using both order-of-magnitude estimates and numerical solutions, we show that the strength of differential rotation required to amplify magnetic field reduces as the aspect ratio of the galactic gas layer increases. Thus, the puzzle of the origin of large-scale magnetic fields appears to be solved. As a result, this class of galaxies provides a new ground for testing our understanding of galactic magnetism. △ Less

Submitted 15 October, 2019; originally announced October 2019.

Comments: Article submitted to Galaxies for a special issue "New Perspectives on Galactic Magnetism"

Turbulent transport coefficients in galactic dynamo simulations using singular value decomposition

Authors: Abhijit B. Bendre, Kandaswamy Subramanian, Detlef Elstner, Oliver Gressel

Abstract: Coherent magnetic fields in disc galaxies are thought to be generated by a large-scale (or mean-field) dynamo operating in their interstellar medium. A key driver of mean magnetic field growth is the turbulent electromotive force (EMF), which represents the influence of correlated small-scale (or fluctuating) velocity and magnetic fields on the mean field. The EMF is usually expressed as a linear… ▽ More Coherent magnetic fields in disc galaxies are thought to be generated by a large-scale (or mean-field) dynamo operating in their interstellar medium. A key driver of mean magnetic field growth is the turbulent electromotive force (EMF), which represents the influence of correlated small-scale (or fluctuating) velocity and magnetic fields on the mean field. The EMF is usually expressed as a linear expansion in the mean magnetic field and its derivatives, with the dynamo tensors as expansion coefficients. Here, we adopt the singular value decomposition (SVD) method to directly measure these turbulent transport coefficients in a simulation of the turbulent interstellar medium that realizes a large-scale dynamo. Specifically, the SVD is used to least-square fit the time series data of the EMF with that of the mean field and its derivatives, to determine these coefficients. We demonstrate that the spatial profiles of the EMF reconstructed from the SVD coefficients match well with that taken directly from the simulation. Also, as a direct test, we use the coefficients to simulate a 1-D mean-field dynamo model and find an overall similarity in the evolution of the mean magnetic field between the dynamo model and the direct simulation. We also compare the results with those which arise using simple regression and the ones obtained previously using the test-field (TF) method, to find reasonable qualitative agreement. Overall, the SVD method provides an effective post-processing tool to determine turbulent transport coefficients from simulations. △ Less

Submitted 18 September, 2019; originally announced September 2019.

Comments: 14 pages, 14 figures

Curriculum based Dropout Discriminator for Domain Adaptation

Authors: Vinod Kumar Kurmi, Vipul Bajaj, Venkatesh K Subramanian, Vinay P Namboodiri

Abstract: Domain adaptation is essential to enable wide usage of deep learning based networks trained using large labeled datasets. Adversarial learning based techniques have shown their utility towards solving this problem using a discriminator that ensures source and target distributions are close. However, here we suggest that rather than using a point estimate, it would be useful if a distribution based… ▽ More Domain adaptation is essential to enable wide usage of deep learning based networks trained using large labeled datasets. Adversarial learning based techniques have shown their utility towards solving this problem using a discriminator that ensures source and target distributions are close. However, here we suggest that rather than using a point estimate, it would be useful if a distribution based discriminator could be used to bridge this gap. This could be achieved using multiple classifiers or using traditional ensemble methods. In contrast, we suggest that a Monte Carlo dropout based ensemble discriminator could suffice to obtain the distribution based discriminator. Specifically, we propose a curriculum based dropout discriminator that gradually increases the variance of the sample based distribution and the corresponding reverse gradients are used to align the source and target feature representations. The detailed results and thorough ablation analysis show that our model outperforms state-of-the-art results. △ Less

Submitted 19 October, 2019; v1 submitted 24 July, 2019; originally announced July 2019.

Comments: BMVC 2019 Accepted, Project Page: https://delta-lab-iitk.github.io/CD3A/

Mean Spectral Normalization of Deep Neural Networks for Embedded Automation

Authors: Anand Krishnamoorthy Subramanian, Nak Young Chong

Abstract: Deep Neural Networks (DNNs) have begun to thrive in the field of automation systems, owing to the recent advancements in standardising various aspects such as architecture, optimization techniques, and regularization. In this paper, we take a step towards a better understanding of Spectral Normalization (SN) and its potential for standardizing regularization of a wider range of Deep Learning model… ▽ More Deep Neural Networks (DNNs) have begun to thrive in the field of automation systems, owing to the recent advancements in standardising various aspects such as architecture, optimization techniques, and regularization. In this paper, we take a step towards a better understanding of Spectral Normalization (SN) and its potential for standardizing regularization of a wider range of Deep Learning models, following an empirical approach. We conduct several experiments to study their training dynamics, in comparison with the ubiquitous Batch Normalization (BN) and show that SN increases the gradient sparsity and controls the gradient variance. Furthermore, we show that SN suffers from a phenomenon, we call the mean-drift effect, which mitigates its performance. We, then, propose a weight reparameterization called as the Mean Spectral Normalization (MSN) to resolve the mean drift, thereby significantly improving the network's performance. Our model performs ~16% faster as compared to BN in practice, and has fewer trainable parameters. We also show the performance of our MSN for small, medium, and large CNNs - 3-layer CNN, VGG7 and DenseNet-BC, respectively - and unsupervised image generation tasks using Generative Adversarial Networks (GANs) to evaluate its applicability for a broad range of embedded automation tasks. △ Less

Submitted 9 July, 2019; originally announced July 2019.

Comments: 8 pagesm IEEE CASE 2019

Efficient quasi-kinematic large-scale dynamo as the small-scale dynamo saturates

Authors: Pallavi Bhat, Kandaswamy Subramanian, Axel Brandenburg

Abstract: Large-scale magnetic fields in stars and galaxies are thought to arise by mean-field dynamo action due to the combined influence of both helical turbulence and shear. Those systems are also highly conducting and the turbulence therein leads to a fluctuation (or small-scale) dynamo which more rapidly amplifies magnetic field fluctuations on the eddy scales and smaller. Will this then interfere with… ▽ More Large-scale magnetic fields in stars and galaxies are thought to arise by mean-field dynamo action due to the combined influence of both helical turbulence and shear. Those systems are also highly conducting and the turbulence therein leads to a fluctuation (or small-scale) dynamo which more rapidly amplifies magnetic field fluctuations on the eddy scales and smaller. Will this then interfere with and suppress the mean (or large-scale) field growth? Using direct numerical simulations of helical turbulence (with and without shear), we identify a novel quasi-kinematic large-scale dynamo which operates as the small-scale dynamo saturates. Thus both dynamos operate efficiently, one after the other, and lead to the generation of significant large-scale fields. △ Less

Submitted 20 May, 2019; originally announced May 2019.

Comments: 8 pages, 11 figures. Comments/feedback/discussions are welcome

Report number: NORDITA-2019-044

From primordial seed magnetic fields to the galactic dynamo

Authors: Kandaswamy Subramanian

Abstract: The origin and maintenance of coherent magnetic fields in the Universe is reviewed with an emphasis on the possible challenges that arise in their theoretical understanding. We begin with the interesting possibility that magnetic fields originated at some level from the early universe. This could be during inflation, the electroweak or the quark-hadron phase transitions. These mechanisms can give… ▽ More The origin and maintenance of coherent magnetic fields in the Universe is reviewed with an emphasis on the possible challenges that arise in their theoretical understanding. We begin with the interesting possibility that magnetic fields originated at some level from the early universe. This could be during inflation, the electroweak or the quark-hadron phase transitions. These mechanisms can give rise to fields which could be strong, but often with much smaller coherence scales than galactic scales. Their subsequent turbulent decay decreases their strength but increases their coherence. We then turn to astrophysical batteries which can generate seed magnetic fields. Here the coherence scale can be large, but the field strength is generally very small. These seed fields need to be further amplified and maintained by a dynamo to explain observed magnetic fields in galaxies. Basic ideas behind both small and large-scale turbulent dynamos are outlined. The small-scale dynamo may help understand the first magnetization of young galaxies, while the large-scale dynamo is important for the generation of fields with scales larger than the stirring scale, as observed in nearby disk galaxies. The current theoretical challenges that turbulent dynamos encounter and their possible resolution are discussed. △ Less

Submitted 9 April, 2019; v1 submitted 9 March, 2019; originally announced March 2019.

Comments: Invited Review Article for "New perspectives on galactic magnetism", Accepted in Galaxies

Journal ref: Galaxies, vol. 7, issue 2, p. 47, 2019

Magnetic Fields in the Universe

Authors: Kandaswamy Subramanian

Abstract: The origin and maintenance of large scale magnetic fields in the Universe is discussed. A popular idea is that cosmic batteries generate seed magnetic fields which were amplified by magnetic induction due to turbulent motions, at times combined with differential rotation. We outline a seed field mechanism, the Biermann battery and then consider some basic ideas behind both small and large-scale tu… ▽ More The origin and maintenance of large scale magnetic fields in the Universe is discussed. A popular idea is that cosmic batteries generate seed magnetic fields which were amplified by magnetic induction due to turbulent motions, at times combined with differential rotation. We outline a seed field mechanism, the Biermann battery and then consider some basic ideas behind both small and large-scale turbulent dynamos. The small-scale dynamo may help understand magnetism in galaxy clusters and young galaxies, while the large-scale dynamo is important for the generation of fields with scales larger than stirring, as observed in nearby disk galaxies. Another intriguing possibility is that magnetic fields originated at some level from the early universe. △ Less

Submitted 10 September, 2018; originally announced September 2018.

Comments: To be published in the M.N. Saha 125th anniversary issue of "Science and Culture"

Probing feedback in high-$z$ galaxies using extended UV luminosity functions

Authors: Saumyadip Samui, Raghunathan Srianand, Kandaswamy Subramanian

Abstract: We fit the recently updated UV luminosity functions (LF) of high-$z$ ($1.5\le z\le8.0$) galaxies using our semi-analytical models of galaxy formation that take into account various feedback processes. In order to reproduce the overall redshift evolution we require the efficiency of converting gas into stars to decrease with decreasing redshift. Even for $z\ge 6$, our models require supernovae (SNe… ▽ More We fit the recently updated UV luminosity functions (LF) of high-$z$ ($1.5\le z\le8.0$) galaxies using our semi-analytical models of galaxy formation that take into account various feedback processes. In order to reproduce the overall redshift evolution we require the efficiency of converting gas into stars to decrease with decreasing redshift. Even for $z\ge 6$, our models require supernovae (SNe) feedback to reproduce the observed LF suggesting the prevalence of galactic winds that could have polluted the inter-galactic medium even at very high redshifts. The observed LF in the low luminosity end for $z<2.5$ shows an upward turn. In our models we reproduce this trend using passively evolving population of galaxies. Measuring stellar mass, age and metallicity of these galaxies using multi-band spectral energy distribution (SED) fitting will place strong constraints on the existence of such galaxies. While Active galactic Nuclei (AGN) feedback is essential to reproduce the LF at high luminosity end for $z<4$, it may not needed in the case of $z\ge 6$. We show that the expected turn around in the LF due to cooling criteria occurs at luminosity much lower than what has been probed with present day observations. With future deep observations, that can measure the LF more accurately, we will be able to distinguish between different modes of SNe feedback and get insights into the physical processes that drive the galaxy evolution. △ Less

Submitted 15 May, 2018; originally announced May 2018.

Comments: 9 pages, 6 figures, 1 table, submitted to MNRAS

Varying the forcing scale in low Prandtl number dynamos

Authors: A. Brandenburg, N. E. L. Haugen, Xiang-Yu Li, K. Subramanian

Abstract: Small-scale dynamos are expected to operate in all astrophysical fluids that are turbulent and electrically conducting, for example the interstellar medium, stellar interiors, and accretion disks, where they may also be affected by or competing with large-scale dynamos. However, the possibility of small-scale dynamos being excited at small and intermediate ratios of viscosity to magnetic diffusivi… ▽ More Small-scale dynamos are expected to operate in all astrophysical fluids that are turbulent and electrically conducting, for example the interstellar medium, stellar interiors, and accretion disks, where they may also be affected by or competing with large-scale dynamos. However, the possibility of small-scale dynamos being excited at small and intermediate ratios of viscosity to magnetic diffusivity (the magnetic Prandtl number) has been debated, and the possibility of them depending on the large-scale forcing wavenumber has been raised. Here we show, using four values of the forcing wavenumber, that the small-scale dynamo does not depend on the scale-separation between the size of the simulation domain and the integral scale of the turbulence, i.e., the forcing scale. Moreover, the spectral bottleneck in turbulence, which has been implied as being responsible for raising the excitation conditions of small-scale dynamos, is found to be invariant under changing the forcing wavenumber. However, when forcing at the lowest few wavenumbers, the effective forcing wavenumber that enters in the definition of the magnetic Reynolds number is found to be about twice the minimum wavenumber of the domain. Our work is relevant to future studies of small-scale dynamos, of which several applications are being discussed. △ Less

Submitted 6 June, 2018; v1 submitted 3 May, 2018; originally announced May 2018.

Comments: 8 pages, 5 figures, MNRAS, resubmitted

Report number: Nordita-2018-038

Journal ref: Mon. Not. Roy. Astron. Soc. 479 (2018) 2827-2833

The generation of helical magnetic field in a viable scenario of Inflationary Magnetogenesis

Authors: Ramkishor Sharma, Kandaswamy Subramanian, T. R. Seshadri

Abstract: We study the generation of helical magnetic fields in a model of inflationary magnetogenesis which is free from the strong coupling and back-reaction problems. To generate helical magnetic fields, we add an $f^2 \tilde{F}^{μν} F_{μν}$ term to the lagrangian of Ratra model. The strong coupling and back-reaction problems are avoided if we take a particular behaviour of coupling function $f$, in whic… ▽ More We study the generation of helical magnetic fields in a model of inflationary magnetogenesis which is free from the strong coupling and back-reaction problems. To generate helical magnetic fields, we add an $f^2 \tilde{F}^{μν} F_{μν}$ term to the lagrangian of Ratra model. The strong coupling and back-reaction problems are avoided if we take a particular behaviour of coupling function $f$, in which $f$ increases during inflation and decreases post inflation to reheating. The generated magnetic field is fully helical and has a blue spectrum, $dρ_B/d\ln k \propto k^4$. This spectrum is obtained when coupling function $f\propto a^2$ during inflation. The scale of reheating in our model has to be lower than $4000$ GeV to avoid back-reaction post inflation. The generated magnetic field spectrum satisfies the $γ$-ray bound for all the possible scales of reheating. The comoving magnetic field strength and its correlation length are $\sim 4 \times 10^{-11} $ G and $70$ kpc respectively, if reheating takes place at 100 GeV. For reheating at the QCD scales of $150$ MeV, the field strength increases to $\sim$ nano gauss, with coherence scale of $0.6$ Mpc. △ Less

Submitted 13 February, 2018; originally announced February 2018.

Comments: 11 pages, Submitted to PRD

Journal ref: Phys. Rev. D 97, 083503 (2018)

Faraday rotation signatures of fluctuation dynamos in young galaxies

Authors: Sharanya Sur, Pallavi Bhat, Kandaswamy Subramanian

Abstract: Observations of Faraday rotation through high-redshift galaxies have revealed that they host coherent magnetic fields that are of comparable strengths to those observed in nearby galaxies. These fields could be generated by fluctuation dynamos. We use idealized numerical simulations of such dynamos in forced compressible turbulence up to rms Mach number of 2.4 to probe the resulting rotation measu… ▽ More Observations of Faraday rotation through high-redshift galaxies have revealed that they host coherent magnetic fields that are of comparable strengths to those observed in nearby galaxies. These fields could be generated by fluctuation dynamos. We use idealized numerical simulations of such dynamos in forced compressible turbulence up to rms Mach number of 2.4 to probe the resulting rotation measure (RM) and the degree of coherence of the magnetic field. We obtain rms values of RM at dynamo saturation of the order of 45 - 55 per cent of the value expected in a model where fields are assumed to be coherent on the forcing scale of turbulence. We show that the dominant contribution to the RM in subsonic and transonic cases comes from the general sea of volume filling fields, rather than from the rarer structures. However, in the supersonic case, strong field regions as well as moderately overdense regions contribute significantly. Our results can account for the observed RMs in young galaxies. △ Less

Submitted 12 February, 2018; v1 submitted 23 November, 2017; originally announced November 2017.

Comments: Published in MNRAS Letters

Journal ref: MNRAS, 475, L72 - L76 (2018)

Intention-Net: Integrating Planning and Deep Learning for Goal-Directed Autonomous Navigation

Authors: Wei Gao, David Hsu, Wee Sun Lee, Shengmei Shen, Karthikk Subramanian

Abstract: How can a delivery robot navigate reliably to a destination in a new office building, with minimal prior information? To tackle this challenge, this paper introduces a two-level hierarchical approach, which integrates model-free deep learning and model-based path planning. At the low level, a neural-network motion controller, called the intention-net, is trained end-to-end to provide robust local… ▽ More How can a delivery robot navigate reliably to a destination in a new office building, with minimal prior information? To tackle this challenge, this paper introduces a two-level hierarchical approach, which integrates model-free deep learning and model-based path planning. At the low level, a neural-network motion controller, called the intention-net, is trained end-to-end to provide robust local navigation. The intention-net maps images from a single monocular camera and "intentions" directly to robot controls. At the high level, a path planner uses a crude map, e.g., a 2-D floor plan, to compute a path from the robot's current location to the goal. The planned path provides intentions to the intention-net. Preliminary experiments suggest that the learned motion controller is robust against perceptual uncertainty and by integrating with a path planner, it generalizes effectively to new environments and goals. △ Less

Submitted 16 October, 2017; v1 submitted 16 October, 2017; originally announced October 2017.

Comments: Published in 1st Annual Conference on Robot Learning (CoRL 2017)

Challenges in Inflationary Magnetogenesis: Constraints from Strong Coupling, Backreaction and the Schwinger Effect

Authors: Ramkishor Sharma, Sandhya Jagannathan, T. R. Seshadri, Kandaswamy Subramanian

Abstract: Models of inflationary magnetogenesis with a coupling to the electromagnetic action of the form $f^2 F_{μν}F^{μν}$, are known to suffer from several problems. These include the strong coupling problem, the back reaction problem and also strong constraints due to Schwinger effect. We propose a model which resolves all these issues. In our model, the coupling function, $f$, grows during inflation an… ▽ More Models of inflationary magnetogenesis with a coupling to the electromagnetic action of the form $f^2 F_{μν}F^{μν}$, are known to suffer from several problems. These include the strong coupling problem, the back reaction problem and also strong constraints due to Schwinger effect. We propose a model which resolves all these issues. In our model, the coupling function, $f$, grows during inflation and transits to a decaying phase post inflation. This evolutionary behaviour is chosen so as to avoid the problem of strong coupling. By assuming a suitable power law form of the coupling function, we can also neglect back reaction effects during inflation. To avoid back reaction post-inflation, we find that the reheating temperature is restricted to be below $ \approx 1.7 \times 10^4$ GeV. The magnetic energy spectrum is predicted to be non-helical and generically blue. The estimated present day magnetic field strength and the corresponding coherence length taking reheating at the QCD epoch(150 MeV) are $ 1.4 \times 10^{-12}$ G and $6.1 \times 10^{-4}$ Mpc, respectively. This is obtained after taking account of nonlinear processing over and above the flux freezing evolution after reheating. If we consider also the possibility of a non-helical inverse cascade, as indicated in direct numerical simulations, the coherence length and the magnetic field strength are even larger. In all cases mentioned above, the magnetic fields generated in our models satisfy the $γ$-ray bound below a certain reheating temperature. △ Less

Submitted 14 October, 2017; v1 submitted 27 August, 2017; originally announced August 2017.

Comments: 13 pages, 4 figures, Published in PRD

Journal ref: Phys. Rev. D 96, 083511 (2017)

Two-Dimensional Pattern Languages

Authors: Henning Fernau, Markus L. Schmid, K. G. Subramanian

Abstract: We introduce several classes of array languages obtained by generalising Angluin's pattern languages to the two-dimensional case. These classes of two-dimensional pattern languages are compared with respect to their expressive power and their closure properties are investigated. We introduce several classes of array languages obtained by generalising Angluin's pattern languages to the two-dimensional case. These classes of two-dimensional pattern languages are compared with respect to their expressive power and their closure properties are investigated. △ Less

Submitted 13 July, 2017; originally announced July 2017.

Efficient cold outflows driven by cosmic rays in high redshift galaxies and their global effects on the IGM

Authors: Saumyadip Samui, Kandaswamy Subramanian, Raghunathan Srianand

Abstract: We present semi-analytical models of galactic outflows in high redshift galaxies driven by both hot thermal gas and non-thermal cosmic rays. Thermal pressure alone may not sustain a large scale outflow in low mass galaxies (i.e $M\sim 10^8$~M$_\odot$), in the presence of supernovae (SNe) feedback with large mass loading. We show that inclusion of cosmic ray pressure allows outflow solutions even i… ▽ More We present semi-analytical models of galactic outflows in high redshift galaxies driven by both hot thermal gas and non-thermal cosmic rays. Thermal pressure alone may not sustain a large scale outflow in low mass galaxies (i.e $M\sim 10^8$~M$_\odot$), in the presence of supernovae (SNe) feedback with large mass loading. We show that inclusion of cosmic ray pressure allows outflow solutions even in these galaxies. In massive galaxies for the same energy efficiency, cosmic ray driven winds can propagate to larger distances compared to pure thermally driven winds. On an average gas in the cosmic ray driven winds has a lower temperature which could aid detecting it through absorption lines in the spectra of background sources. Using our constrained semi-analytical models of galaxy formation (that explains the observed UV luminosity functions of galaxies) we study the influence of cosmic ray driven winds on the properties of the intergalactic medium (IGM) at different redshifts. In particular, we study the volume filling factor, average metallicity, cosmic ray and magnetic field energy densities for models invoking atomic cooled and molecular cooled halos. We show that the cosmic rays in the IGM could have enough energy that can be transferred to the thermal gas in presence of magnetic fields to influence the thermal history of the intergalactic medium. The significant volume filling and resulting strength of IGM magnetic fields can also account for recent $γ$-ray observations of blazars. △ Less

Submitted 31 January, 2018; v1 submitted 6 June, 2017; originally announced June 2017.

Comments: 16 Pages, 9 figures, Revised version, Accepted for publication in MNRAS

Navigating Occluded Intersections with Autonomous Vehicles using Deep Reinforcement Learning

Authors: David Isele, Reza Rahimi, Akansel Cosgun, Kaushik Subramanian, Kikuo Fujimura

Abstract: Providing an efficient strategy to navigate safely through unsignaled intersections is a difficult task that requires determining the intent of other drivers. We explore the effectiveness of Deep Reinforcement Learning to handle intersection problems. Using recent advances in Deep RL, we are able to learn policies that surpass the performance of a commonly-used heuristic approach in several metric… ▽ More Providing an efficient strategy to navigate safely through unsignaled intersections is a difficult task that requires determining the intent of other drivers. We explore the effectiveness of Deep Reinforcement Learning to handle intersection problems. Using recent advances in Deep RL, we are able to learn policies that surpass the performance of a commonly-used heuristic approach in several metrics including task completion time and goal success rate and have limited ability to generalize. We then explore a system's ability to learn active sensing behaviors to enable navigating safely in the case of occlusions. Our analysis, provides insight into the intersection handling problem, the solutions learned by the network point out several shortcomings of current rule-based methods, and the failures of our current deep reinforcement learning system point to future research directions. △ Less

Submitted 26 February, 2018; v1 submitted 2 May, 2017; originally announced May 2017.

Comments: IEEE International Conference on Robotics and Automation (ICRA 2018)

Understanding the non-linear clustering of high redshift galaxies

Authors: Charles Jose, Carlton M. Baugh, Cedric G. Lacey, Kandaswamy Subramanian

Abstract: We incorporate the non-linear clustering of dark matter halos, as modelled by Jose et al. (2016) into the halo model to better understand the clustering of Lyman break galaxies (LBGs) in the redshift range $z=3-5$. We find that, with this change, the predicted LBG clustering increases significantly on quasi-linear scales ($0.1 \leq r\,/\,h^{-1} \,{\rm Mpc} \leq 10$) compared to that in the linear… ▽ More We incorporate the non-linear clustering of dark matter halos, as modelled by Jose et al. (2016) into the halo model to better understand the clustering of Lyman break galaxies (LBGs) in the redshift range $z=3-5$. We find that, with this change, the predicted LBG clustering increases significantly on quasi-linear scales ($0.1 \leq r\,/\,h^{-1} \,{\rm Mpc} \leq 10$) compared to that in the linear halo bias model. This in turn results in an increase in the clustering of LBGs by an order of magnitude on angular scales $5" \leq θ\leq 100"$. Remarkably, the predictions of our new model on the whole remove the systematic discrepancy between the linear halo bias predictions and the observations. The correlation length and large scale galaxy bias of LBGs are found to be significantly higher in the non-linear halo bias model than in the linear halo bias model. The resulting two-point correlation function retains an approximate power-law form in contrast with that computed using the linear halo bias theory. We also find that the non-linear clustering of LBGs increases with increasing luminosity and redshift. Our work emphasizes the importance of using non-linear halo bias in order to model the clustering of high-z galaxies to probe the physics of galaxy formation and extract cosmological parameters reliably. △ Less

Submitted 6 June, 2017; v1 submitted 2 February, 2017; originally announced February 2017.

Comments: Accepted for publication in MNRAS

Probing magnetic fields with Square Kilometre array and its precursors

Authors: Subhashis Roy, Sharanya Sur, Kandaswamy Subramanian, Arun Mangalam, TR Seshadri, Hum Chand

Abstract: Origin of magnetic fields, its structure and effects on dynamical processes in stars to galaxies are not well understood. Lack of a direct probe has hampered its study. The first phase of Square Kilometre Array (SKA-I), will have more than an order of magnitude higher sensitivity than existing radio telescopes. In this contribution, we discuss specific science cases that are of interest to the Ind… ▽ More Origin of magnetic fields, its structure and effects on dynamical processes in stars to galaxies are not well understood. Lack of a direct probe has hampered its study. The first phase of Square Kilometre Array (SKA-I), will have more than an order of magnitude higher sensitivity than existing radio telescopes. In this contribution, we discuss specific science cases that are of interest to the Indian community concerned with astrophysical turbulence and magnetic fields. The SKA-I will allow observations of a large number of background sources with detectable polarisation and measure their Faraday depths (FDs) through the Milky Way, other galaxies and their circum-galactic medium. This will probe line-of-sight magnetic fields in these objects well and provide field configurations. Detailed comparison of observational data with models which consider various processes giving rise to field amplification and maintenance will then be possible. Such observations will also provide the coherence scale of the fields and measure its random component. Measuring the random component is important to characterise turbulence in the medium. Observations of FDs with redshift will provide important information on magnetic field evolution as a function of redshift. The background sources could also be used to probe magnetic fields and its coherent scale in galaxy clusters and in bridges formed between interacting galaxies. Other than FDs, sensitive observations of synchrotron emission from galaxies will provide complimentary information on their magnetic field strengths in the sky plane. The core shift measurements of AGNs can provide more precise measurements of magnetic field very close (<pc) to the black hole and its evolution. The low band of SKA-I will also be useful to study circularly polarized emission from Sun and comparing various models of field configurations with observations. △ Less

Submitted 21 November, 2016; originally announced November 2016.

Comments: To appear in Journal of Astrophysics and Astronomy (JOAA) special issue on "Science with the SKA: an Indian perspective"

Evolution of the magnetorotational instability on initially tangled magnetic fields

Authors: Pallavi Bhat, Fatima Ebrahimi, Eric G. Blackman, Kandaswamy Subramanian

Abstract: The initial magnetic field of previous magnetorotational instability (MRI) simulations has always included a significant system-scale component, even if stochastic. However, it is of conceptual and practical interest to assess whether the MRI can grow when the initial field is turbulent. The ubiquitous presence of turbulent or random flows in astrophysical plasmas generically leads to a small-scal… ▽ More The initial magnetic field of previous magnetorotational instability (MRI) simulations has always included a significant system-scale component, even if stochastic. However, it is of conceptual and practical interest to assess whether the MRI can grow when the initial field is turbulent. The ubiquitous presence of turbulent or random flows in astrophysical plasmas generically leads to a small-scale dynamo (SSD), which would provide initial seed turbulent velocity and magnetic fields in the plasma that becomes an accretion disc. Can the MRI grow from these more realistic initial conditions? To address this we supply a standard shearing box with isotropically forced SSD generated magnetic and velocity fields as initial conditions, and remove the forcing. We find that if the initially supplied fields are too weak or too incoherent, they decay from the initial turbulent cascade faster than they can grow via the MRI. When the initially supplied fields are sufficient to allow MRI growth and sustenance, the saturated stresses, large-scale fields, and power spectra match those of the standard zero net flux MRI simulation with an initial large scale vertical field. △ Less

Submitted 11 October, 2017; v1 submitted 21 September, 2016; originally announced September 2016.

Comments: Published in MNRAS, 6 pages, 5 figures

Measuring correlations in non-separable vector beams using projective measurements

Authors: Keerthan Subramanian, Nirmal K. Viswanathan

Abstract: Doubts regarding the completeness of quantum mechanics as raised by Einstein, Podolsky and Rosen(EPR) have predominantly been resolved by resorting to a measurement of correlations between entangled photons which clearly demonstrate violation of Bell's inequality. This article is an attempt to reconcile incompatibility of hidden variable theories with reality by demonstrating experimentally a viol… ▽ More Doubts regarding the completeness of quantum mechanics as raised by Einstein, Podolsky and Rosen(EPR) have predominantly been resolved by resorting to a measurement of correlations between entangled photons which clearly demonstrate violation of Bell's inequality. This article is an attempt to reconcile incompatibility of hidden variable theories with reality by demonstrating experimentally a violation of Bell's inequality in locally correlated systems whose two degrees of freedom, the spin and orbital angular momentum, are maximally correlated. To this end we propose and demonstrate a linear, achromatic modified Sagnac interferometer to project orbital angular momentum states which we combine with spin projections to measure correlations. △ Less

Submitted 4 April, 2017; v1 submitted 22 February, 2016; originally announced February 2016.

Comments: 18 pages, 6 figures

Journal ref: Optics Communications 399(2017) 45-51