arXiv subjects

Ankit Singh

Publications and source records attributed to Ankit Singh.

At least 37 records · Page 2Linked to original sources

arXiv2024

Optimizing growth performance of Abelmoschus esculentus (L.) via synergistic effects of biogenic Cu/Ni/Co oxide nanoparticles in conjunction with rice straw and pressmud based vermicompost

This study is a continuation of previous work, which highlights the nutrient enhancement by using rice straw (RS) and pressmud (PM) on vermicomposting. Herein, we demonstrate the significant impact of Moringa oleifera derived Cu/Ni/Co oxide nanoparticles (TmONs) in conjunction with these vermicompost on the growth performance of Abelmoschus esculentus. Vermicompost produced under various combinations (T0, cow dung (CD) only; T1, 1CD:1RS; T2, 1CD:1PM, and T3, 1CD:1RS:1PM) were further enriched by blending with biogenic nanoparticles. This strategic combination enhances the nutritional composition of the vermicompost, contributing to its overall effectiveness in promoting plant growth and health. Various analytical techniques, including FTIR, XRD, XPS, FESEM-EDX, TEM, and ICP-OES, were employed for comprehensive characterization. The synthesized TmONs with sizes ranging from 13 to 54 nm exhibited distinct CuO, NiO, and CoO phases. The vermicompost blended TmONs demonstrated significant improvements (P < 0.05) in seed germination (167%), coefficient velocity (67%), and vigour index (95%), while reducing the mean germination time by 41% for A. esculentus compared to the control group. The plant culture group nT3 (T3 + TmONs) showed the best growth performance. Furthermore, trace element concentrations in both soil and plant leaves were found to be below the maximum permissible limits set by WHO (1996). This investigation extends the understanding of the role played by these nanoparticles in fostering optimal conditions for plant growth and development as these micronutrients are essential components for several plant enzymes.

q-bio.OT↗

arXiv2024

Falcon2-11B Technical Report

We introduce Falcon2-11B, a foundation model trained on over five trillion tokens, and its multimodal counterpart, Falcon2-11B-vlm, which is a vision-to-text model. We report our findings during the training of the Falcon2-11B which follows a multi-stage approach where the early stages are distinguished by their context length and a final stage where we use a curated, high-quality dataset. Additionally, we report the effect of doubling the batch size mid-training and how training loss spikes are affected by the learning rate. The downstream performance of the foundation model is evaluated on established benchmarks, including multilingual and code datasets. The foundation model shows strong generalization across all the tasks which makes it suitable for downstream finetuning use cases. For the vision language model, we report the performance on several benchmarks and show that our model achieves a higher average score compared to open-source models of similar size. The model weights and code of both Falcon2-11B and Falcon2-11B-vlm are made available under a permissive license.

cs.CL↗

arXiv2024

The Environmental Dependence of the Stellar Mass - Gas Metallicity Relation in Horizon Run 5

Metallicity offers a unique window into the baryonic history of the cosmos, being instrumental in probing evolutionary processes in galaxies between different cosmic environments. We aim to quantify the contribution of these environments to the scatter in the mass-metallicity relation (MZR) of galaxies. By analysing the galaxy distribution within the cosmic skeleton of the Horizon Run 5 cosmological hydrodynamical simulation at redshift $z = 0.625$, computed using a careful calibration of the T-ReX filament finder, we identify galaxies within three main environments: nodes, filaments and voids. We also classify galaxies based on the dynamical state of the clusters and the length of the filaments in which they reside. We find that the cosmic environment significantly contributes to the scatter in the MZR; in particular, both the gas metallicity and its average relative standard deviation increase when considering denser large-scale environments. The difference in the average metallicity between galaxies within relaxed and unrelaxed clusters is $\approx 0.1 \text{ dex}$, with both populations displaying positive residuals, $δZ_{g}$, from the averaged MZR. Moreover, the difference in metallicity between node and void galaxies accounts for $\approx 0.14 \, \text{dex}$ in the scatter of the MZR at stellar mass $M_{\star} \approx 10^{9.35}\,\text{M}_{\odot}$. Finally, both the average [O/Fe] in the gas and the galaxy gas fraction decrease when moving to higher large-scale densities in the simulation, suggesting that the cores of cosmic environments host, on average, older and more massive galaxies, whose enrichment is affected by a larger number of Type Ia Supernova events.

astro-ph.GA↗

arXiv2024

ViSpeR: Multilingual Audio-Visual Speech Recognition

This work presents an extensive and detailed study on Audio-Visual Speech Recognition (AVSR) for five widely spoken languages: Chinese, Spanish, English, Arabic, and French. We have collected large-scale datasets for each language except for English, and have engaged in the training of supervised learning models. Our model, ViSpeR, is trained in a multi-lingual setting, resulting in competitive performance on newly established benchmarks for each language. The datasets and models are released to the community with an aim to serve as a foundation for triggering and feeding further research work and exploration on Audio-Visual Speech Recognition, an increasingly important area of research. Code available at \href{https://github.com/YasserdahouML/visper}{https://github.com/YasserdahouML/visper}.

cs.CL↗

arXiv2024

Study of dependence of ram pressure stripping on the orbital parameters of the galaxies

Comprehensive observations of galaxy clusters suggest that gas deficiency in the galaxies could be due to ram pressure stripping due to the high-pressure intra-cluster medium acting on the galactic discs. The presence of gas in galaxies is essential for star formation. The net force due to ram pressure is dependent on the ambient medium and the orbit followed by the galaxy as it moves past the cluster medium. The current work deals with the effect of non-radial orbits of galaxies and the inclination of the disc plane of galaxies with the orbital plane on the mass of gas removed due to ram pressure. This gives a realistic approach to understanding the process of ram pressure stripping. The orbital parameters are extracted from EAGLE simulation data set along with the mass distribution of the galaxies. The analytical model proposed by Singh et. al. (2019) is modified appropriately to include the effect of the inclination angle. The non-radial orbits and infalling galaxies not being face-on decrease the amount of gas removed. Moreover, the inclination angle has a pronounced effect on the stripping of gas in low-mass galaxies as compared to high-mass galaxies with similar inclinations. The results show that the efficiency of the ram pressure stripping can be much lower in some cases, and hence gas in infalling galaxies can survive for much longer than expected from a simple analysis.

astro-ph.GA↗

arXiv2024

Spatial Distribution of Intracluster Light versus Dark Matter in Horizon Run 5

One intriguing approach for studying the dynamical evolution of galaxy clusters is to compare the spatial distributions among various components, such as dark matter, member galaxies, gas, and intracluster light (ICL). Utilizing the recently introduced Weighted Overlap Coefficient (WOC) \citep{2022ApJS..261...28Y}, we analyze the spatial distributions of components within 174 galaxy clusters ($M_{\rm tot}> 5 \times 10^{13} M_{\odot}$, $z=0.625$) at varying dynamical states in the cosmological hydrodynamical simulation Horizon Run 5. We observe that the distributions of gas and the combination of ICL with the brightest cluster galaxy (BCG) closely resembles the dark matter distribution, particularly in more relaxed clusters, characterized by the half-mass epoch. The similarity in spatial distribution between dark matter and BCG+ICL mimics the changes in the dynamical state of clusters during a major merger. Notably, at redshifts $>$ 1, BCG+ICL traced dark matter more accurately than the gas. Additionally, we examined the one-dimensional radial profiles of each component, which show that the BCG+ICL is a sensitive component revealing the dynamical state of clusters. We propose a new method that can approximately recover the dark matter profile by scaling the BCG+ICL radial profile. Furthermore, we find a recipe for tracing dark matter in unrelaxed clusters by including the most massive satellite galaxies together with BCG+ICL distribution. Combining the BCG+ICL and the gas distribution enhances the dark matter tracing ability. Our results imply that the BCG+ICL distribution is an effective tracer for the dark matter distribution, and the similarity of spatial distribution may be a useful probe of the dynamical state of a cluster.

astro-ph.CO↗

arXiv2023

On permutation symmetries in Bayesian neural network posteriors: a variational perspective

The elusive nature of gradient-based optimization in neural networks is tied to their loss landscape geometry, which is poorly understood. However recent work has brought solid evidence that there is essentially no loss barrier between the local solutions of gradient descent, once accounting for weight-permutations that leave the network's computation unchanged. This raises questions for approximate inference in Bayesian neural networks (BNNs), where we are interested in marginalizing over multiple points in the loss landscape. In this work, we first extend the formalism of marginalized loss barrier and solution interpolation to BNNs, before proposing a matching algorithm to search for linearly connected solutions. This is achieved by aligning the distributions of two independent approximate Bayesian solutions with respect to permutation matrices. We build on the results of Ainsworth et al. (2023), reframing the problem as a combinatorial optimization one, using an approximation to the sum of bilinear assignment problem. We then experiment on a variety of architectures and datasets, finding nearly zero marginalized loss barriers for linearly connected solutions.

stat.ML↗

arXiv2023

On the Effects of Local Environment on Active Galactic Nucleus (AGN) in the Horizon Run 5 Simulation

We use the Horizon Run 5 cosmological simulation to study the effect of galaxy intrinsic properties and the local environment on AGNs characterized by their threshold of the accretion rate. We select galaxies in the stellar mass range $10^{9.5} \le M^{}{*}/M^{}{\odot} \le 10^{10.5}$ in the snapshot at redshift $z$=0.625. Among various intrinsic properties, we find that the star formation rate of the host galaxy is most correlated to the AGN activity. To quantify the environment, we use background galaxy number density (large-scale environment) and distance and morphological type of the nearest neighbors (small-scale environment), and study their relative effects on the AGN properties. We find that, compared to the background density, the nearest neighbor environment is the dominant quantity determining the bolometric luminosity, star formation rate, and kinematic properties of AGNs and better dictates the gas mass of the host galaxy. We show that the cold gas content in the host galaxies is crucial in triggering AGN activity. However, when the nearest neighbor environment effects start to act at the neighbor distance of less than about half the virial radius of the neighbor, the neighbor environmental effects are the most dominant factor for quasar activity.

astro-ph.GA↗

arXiv2021

CLDA: Contrastive Learning for Semi-Supervised Domain Adaptation

Unsupervised Domain Adaptation (UDA) aims to align the labeled source distribution with the unlabeled target distribution to obtain domain invariant predictive models. However, the application of well-known UDA approaches does not generalize well in Semi-Supervised Domain Adaptation (SSDA) scenarios where few labeled samples from the target domain are available. In this paper, we propose a simple Contrastive Learning framework for semi-supervised Domain Adaptation (CLDA) that attempts to bridge the intra-domain gap between the labeled and unlabeled target distributions and inter-domain gap between source and unlabeled target distribution in SSDA. We suggest employing class-wise contrastive learning to reduce the inter-domain gap and instance-level contrastive alignment between the original (input image) and strongly augmented unlabeled target images to minimize the intra-domain discrepancy. We have shown empirically that both of these modules complement each other to achieve superior performance. Experiments on three well-known domain adaptation benchmark datasets namely DomainNet, Office-Home, and Office31 demonstrate the effectiveness of our approach. CLDA achieves state-of-the-art results on all the above datasets.

cs.CV↗

arXiv2021

Effects of Head-locked Augmented Reality on User's performance and perceived workload

An augmented reality (AR) environment includes a set of digital elements with which the users interact while performing certain tasks. Recent AR head-mounted displays allow users to select how these elements are presented. However, few studies have been conducted to examine the effect of the way of presenting augmented content on user performance and workload. This study aims to evaluate two methods of presenting augmented content - world-locked and head-locked modes in a data entry task. A total of eighteen participants performed the data entry task in this study. The effectiveness of each mode is evaluated in terms of task performance, muscle activity, perceived workload, and usability. The results show that the task completion time is shorter and the typing speed is significantly faster in the head-locked mode while the world-locked mode achieved higher scores in terms of preference. The findings of this study can be applied to AR user interfaces to improve content presentation and enhance the user experience.

cs.HC↗

arXiv2021

Semi-Supervised Action Recognition with Temporal Contrastive Learning

Learning to recognize actions from only a handful of labeled videos is a challenging problem due to the scarcity of tediously collected activity labels. We approach this problem by learning a two-pathway temporal contrastive model using unlabeled videos at two different speeds leveraging the fact that changing video speed does not change an action. Specifically, we propose to maximize the similarity between encoded representations of the same video at two different speeds as well as minimize the similarity between different videos played at different speeds. This way we use the rich supervisory information in terms of `time' that is present in otherwise unsupervised pool of videos. With this simple yet effective strategy of manipulating video playback rates, we considerably outperform video extensions of sophisticated state-of-the-art semi-supervised image recognition methods across multiple diverse benchmark datasets and network architectures. Interestingly, our proposed approach benefits from out-of-domain unlabeled videos showing generalization and robustness. We also perform rigorous ablations and analysis to validate our approach. Project page: https://cvir.github.io/TCL/.

cs.CV↗

arXiv2021

How attractive and repulsive interactions affect structure ordering and dynamics of glass-forming liquids?

The theory developed in our previous papers is applied in this paper to investigate the dependence of slowing down of dynamics of glass-forming liquids on the attractive and repulsive parts of intermolecular interactions. Through an extensive comparison of the behavior of a Lennard-Jones glass-forming liquid and that of its WCA reduction to a model with truncated pair potential without attractive tail, we demonstrate why the two systems exhibit very different dynamics despite having nearly identical pair correlation functions. In particular, we show that local structures characterized by number of mobile and immobile particles around a central particle markedly differ in the two systems at densities and temperatures where their dynamics show large difference and nearly identical where dynamics nearly overlap. We also show how the parameter ψ(T ) that measures the role of fluctuations embedded in the system on size of the cooperatively reorganizing cluster (CRC) and the crossover temperature T a depend on the intermolecular interactions. These parameters stemming from the intermolecular interactions characterize the temperature and density dependence of structural relaxation time τ α. The quantitative and qualitative agreements found with simulation results for the two systems suggest that our theory brings out the underlying features that determine dynamics of glass-forming liquids.

cond-mat.soft↗

arXiv2020

Emergence of cooperatively reorganizing cluster and super-Arrhenius dynamics of fragile supercooled liquids

In this paper we develop a theory to calculate the structural relaxation time τ α of fragile su percooled liquids. Using the information of the configurational entropy and structure we calculate the number of dynamically free, metastable, and stable neighbors around a central particle. In supercooled liquids the cooperatively reorganizing clusters (CRCs) in which the stable neighbors form stable nonchemical bonds with the central particle emerge. For an event of relaxation to take place these bonds have to reorganize irreversibly; the energy involved in the processes is the effective activation energy of relaxation. The theory brings forth a temperature T a and a temper ature dependent parameter ψ(T ) which characterize slowing down of dynamics on cooling. It is shown that the value of ψ(T ) is equal to 1 for T > T a indicating that the underlying microscopic mechanism of relaxation is dominated by the entropy driven processes while for T < T a , ψ(T ) decreases on cooling indicating the emergence of the energy driven processes. This crossover of ψ(T ) from high to low temperatures explains the crossover seen in τ α . The dynamics of systems that may have similar static structure but very different dynamics can be understood in terms of ψ(T ). We present results for the Kob-Anderson model for three densities and show that the calculated values of τ α are in excellent agreement with simulation values for all densities. We also show that when ψ(T ), τ α and other quantities are plotted as a function of T /T a (or T a /T ) the data collapse on master curves.

cond-mat.soft↗

arXiv2020

Mitigating Dataset Imbalance via Joint Generation and Classification

Supervised deep learning methods are enjoying enormous success in many practical applications of computer vision and have the potential to revolutionize robotics. However, the marked performance degradation to biases and imbalanced data questions the reliability of these methods. In this work we address these questions from the perspective of dataset imbalance resulting out of severe under-representation of annotated training data for certain classes and its effect on both deep classification and generation methods. We introduce a joint dataset repairment strategy by combining a neural network classifier with Generative Adversarial Networks (GAN) that makes up for the deficit of training examples from the under-representated class by producing additional training examples. We show that the combined training helps to improve the robustness of both the classifier and the GAN against severe class imbalance. We show the effectiveness of our proposed approach on three very different datasets with different degrees of imbalance in them. The code is available at https://github.com/AadSah/ImbalanceCycleGAN .

cs.CV↗

arXiv2020

Study of galaxies on large-scale filaments in simulations

We use data from the Evolution and Assembly of GaLaxies in their Environment (EAGLE) cosmological simulation to study properties of galaxies in the cosmic web. Galaxies become more redder and form stars at a lower rate relative to their counterparts further away from the cylindrical axis of the large-scale filaments. These trends are particularly strong for galaxies with $M_*/M_{\odot}\lesssim10^{10}$. We also find that at distances $<0.5$ Mpc from the spine of the filaments, the median gas and stellar mass fraction in filament galaxies rises sharply with decreasing distance from the spine of the filament. These results, together with matching trends in the SFR/$M_*$ and the $g-r$ colour of filament galaxies suggest that (i) the intrafilamentary gas condenses into the filament galaxies thus fuelling star formation in them, and (ii) increased number density of galaxies closer to the central axis of the filament enhances the rate of gravitational interactions among filament galaxies closer to the spine.

astro-ph.GA↗

arXiv2019

Ram pressure stripping: An analytical approach

We use an analytical approach to study ram pressure stripping with simple models for discs and halo gas distribution to study the phenomena in cluster, group and galaxy halos. We also study variations with galaxy properties and redshift. In each case we model the worst case scenario (i.e., maximum effect due to ram pressure). We show that there is little variation in the worst case scenario with redshift. We find that gas discs in galaxies with a higher spin parameter get stripped sooner than galaxies with a smaller spin parameter. Galaxies in cluster halos get stripped of gas more effciently as compared to group and galaxy halos: this is due to the higher infall speed and a higher density of gas in the ICM due to a greater retention of baryons. We comment on the limitations of our model and situations where a signiffcant amount of gas may be retained in galaxy disc and also give an illustration for the same. Lastly, we discuss implications for star formation in galaxies as these fall into halos.

astro-ph.GA↗

arXiv2019

Understanding the phenomenon of viscous slowing down of glass-forming liquids from the static pair correlation function

A theory which uses data of the static pair-correlation function is developed to calculate quantities associated with the viscous slowing down of supercooled liquids. We calculate value of the energy fluctuations that determine the number of stable bonds a particle forms with neighbors from data of the structural relaxation time. The number of bonds and the activation energy for relaxation are shown to increase sharply in a narrow temperature range close to the glass temperature. The configurational entropy calculated from values of the configurational fluctuations is found in good agreement with the value determined from simulations.

cond-mat.stat-mech↗

arXiv2019

Interdigitated flexible supercapacitor using activated carbon synthesized from biomass for wearable energy storage

We have developed a flexible, interdigitated supercapacitor with high energy storing capacity for wearable, flexible electronic application. Locally obtained low cost biomass (banana peel) was impregnated with KOH under high temperature under an inert atmosphere to synthesize activated carbon with significant Brunauer Emmett Teller surface area of 62.03m2/g. The supercapacitor was fabricated by screen printing interdigitated current collector of conducting silver ink on a thin flexible PET substrate and subsequent deposition of activated carbon and drop casting of gel electrolyte. Fabricated supercapacitor exhibits high capacitance of 33.18 mF/cm2 at 1mV/s scan rate and 20.12 mF/cm2 at a discharge current of 1mA and high energy density of 5.87 micro Wh/cm2. The developed flexible supercapacitor retains its energy storing capacity (90%) over several cycles of mechanical bending and repetitive electronic cycling tests (5000 cycles). Locally available biomass based activated carbon and low cost screen printing technique can be used for large scale fabrication of supercapacitor. The flexible supercapacitor demonstrates high energy storing capacity and mechanical durability through multiple bending and charging and discharging through LED. Therefore, it can be used for further developing integrated wearable and printed electronic devices.

physics.app-ph↗