arXiv subjects

Akash Gupta

Publications and source records attributed to Akash Gupta.

At least 37 records · Page 2Linked to original sources

arXiv2022

UltraMNIST Classification: A Benchmark to Train CNNs for Very Large Images

Convolutional neural network (CNN) approaches available in the current literature are designed to work primarily with low-resolution images. When applied on very large images, challenges related to GPU memory, smaller receptive field than needed for semantic correspondence and the need to incorporate multi-scale features arise. The resolution of input images can be reduced, however, with significant loss of critical information. Based on the outlined issues, we introduce a novel research problem of training CNN models for very large images, and present 'UltraMNIST dataset', a simple yet representative benchmark dataset for this task. UltraMNIST has been designed using the popular MNIST digits with additional levels of complexity added to replicate well the challenges of real-world problems. We present two variants of the problem: 'UltraMNIST classification' and 'Budget-aware UltraMNIST classification'. The standard UltraMNIST classification benchmark is intended to facilitate the development of novel CNN training methods that make the effective use of the best available GPU resources. The budget-aware variant is intended to promote development of methods that work under constrained GPU memory. For the development of competitive solutions, we present several baseline models for the standard benchmark and its budget-aware variant. We study the effect of reducing resolution on the performance and present results for baseline models involving pretrained backbones from among the popular state-of-the-art models. Finally, with the presented benchmark dataset and the baselines, we hope to pave the ground for a new generation of CNN methods suitable for handling large images in an efficient and resource-light manner.

cs.CV↗

arXiv2022

A-ACT: Action Anticipation through Cycle Transformations

While action anticipation has garnered a lot of research interest recently, most of the works focus on anticipating future action directly through observed visual cues only. In this work, we take a step back to analyze how the human capability to anticipate the future can be transferred to machine learning algorithms. To incorporate this ability in intelligent systems a question worth pondering upon is how exactly do we anticipate? Is it by anticipating future actions from past experiences? Or is it by simulating possible scenarios based on cues from the present? A recent study on human psychology explains that, in anticipating an occurrence, the human brain counts on both systems. In this work, we study the impact of each system for the task of action anticipation and introduce a paradigm to integrate them in a learning framework. We believe that intelligent systems designed by leveraging the psychological anticipation models will do a more nuanced job at the task of human action prediction. Furthermore, we introduce cyclic transformation in the temporal dimension in feature and semantic label space to instill the human ability of reasoning of past actions based on the predicted future. Experiments on Epic-Kitchen, Breakfast, and 50Salads dataset demonstrate that the action anticipation model learned using a combination of the two systems along with the cycle transformation performs favorably against various state-of-the-art approaches.

cs.CV↗

arXiv2021

Classification of histopathology images using ConvNets to detect Lupus Nephritis

Systemic lupus erythematosus (SLE) is an autoimmune disease in which the immune system of the patient starts attacking healthy tissues of the body. Lupus Nephritis (LN) refers to the inflammation of kidney tissues resulting in renal failure due to these attacks. The International Society of Nephrology/Renal Pathology Society (ISN/RPS) has released a classification system based on various patterns observed during renal injury in SLE. Traditional methods require meticulous pathological assessment of the renal biopsy and are time-consuming. Recently, computational techniques have helped to alleviate this issue by using virtual microscopy or Whole Slide Imaging (WSI). With the use of deep learning and modern computer vision techniques, we propose a pipeline that is able to automate the process of 1) detection of various glomeruli patterns present in these whole slide images and 2) classification of each image using the extracted glomeruli features.

eess.IV↗

arXiv2021

Photoevaporation vs. core-powered mass-loss: model comparison with the 3D radius gap

The EUV/X-ray photoevaporation and core-powered mass-loss models are both capable of reproducing the bimodality in the sizes of small, close-in exoplanets observed by the \textit{Kepler} space mission, often referred to as the "radius gap". However, it is unclear which of these two mechanisms dominates the atmospheric mass-loss which is likely sculpting the radius gap. In this work, we propose a new method of differentiating between the two models, which relies on analysing the radius gap in 3D parameter space. Using models for both mechanisms, and by performing synthetic transit surveys we predict the size and characteristics of a survey capable of discriminating between the two models. We find that a survey of $\gtrsim 5000$ planets, with a wide range in stellar mass and measurement uncertainties at a $\lesssim 5\%$ level is sufficient. Our methodology is robust against moderate false positive contamination of $\lesssim 10\%$. We perform our analysis on two surveys (which do not satisfy our requirements): the \textit{California Kepler Survey} and the \textit{Gaia-Kepler Survey} and find, unsurprisingly, that both data-sets are consistent with either model. We propose a hypothesis test to be performed on future surveys which can robustly ascertain which of the two mechanisms formed the radius gap, provided one dominates over the other.

astro-ph.EP↗

arXiv2021

APObind: A Dataset of Ligand Unbound Protein Conformations for Machine Learning Applications in De Novo Drug Design

Protein-ligand complex structures have been utilised to design benchmark machine learning methods that perform important tasks related to drug design such as receptor binding site detection, small molecule docking and binding affinity prediction. However, these methods are usually trained on only ligand bound (or holo) conformations of the protein and therefore are not guaranteed to perform well when the protein structure is in its native unbound conformation (or apo), which is usually the conformation available for a newly identified receptor. A primary reason for this is that the local structure of the binding site usually changes upon ligand binding. To facilitate solutions for this problem, we propose a dataset called APObind that aims to provide apo conformations of proteins present in the PDBbind dataset, a popular dataset used in drug design. Furthermore, we explore the performance of methods specific to three use cases on this dataset, through which, the importance of validating them on the APObind dataset is demonstrated.

q-bio.BM↗

arXiv2021

Ada-VSR: Adaptive Video Super-Resolution with Meta-Learning

Most of the existing works in supervised spatio-temporal video super-resolution (STVSR) heavily rely on a large-scale external dataset consisting of paired low-resolution low-frame rate (LR-LFR)and high-resolution high-frame-rate (HR-HFR) videos. Despite their remarkable performance, these methods make a prior assumption that the low-resolution video is obtained by down-scaling the high-resolution video using a known degradation kernel, which does not hold in practical settings. Another problem with these methods is that they cannot exploit instance-specific internal information of video at testing time. Recently, deep internal learning approaches have gained attention due to their ability to utilize the instance-specific statistics of a video. However, these methods have a large inference time as they require thousands of gradient updates to learn the intrinsic structure of the data. In this work, we presentAdaptiveVideoSuper-Resolution (Ada-VSR) which leverages external, as well as internal, information through meta-transfer learning and internal learning, respectively. Specifically, meta-learning is employed to obtain adaptive parameters, using a large-scale external dataset, that can adapt quickly to the novel condition (degradation model) of the given test video during the internal learning task, thereby exploiting external and internal information of a video for super-resolution. The model trained using our approach can quickly adapt to a specific video condition with only a few gradient updates, which reduces the inference time significantly. Extensive experiments on standard datasets demonstrate that our method performs favorably against various state-of-the-art approaches.

eess.IV↗

arXiv2021

Deep Quantized Representation for Enhanced Reconstruction

While machine learning approaches have shown remarkable performance in biomedical image analysis, most of these methods rely on high-quality and accurate imaging data. However, collecting such data requires intensive and careful manual effort. One of the major challenges in imaging the Shoot Apical Meristem (SAM) of Arabidopsis thaliana, is that the deeper slices in the z-stack suffer from different perpetual quality-related problems like poor contrast and blurring. These quality-related issues often lead to the disposal of the painstakingly collected data with little to no control on quality while collecting the data. Therefore, it becomes necessary to employ and design techniques that can enhance the images to make them more suitable for further analysis. In this paper, we propose a data-driven Deep Quantized Latent Representation (DQLR) methodology for high-quality image reconstruction in the Shoot Apical Meristem (SAM) of Arabidopsis thaliana. Our proposed framework utilizes multiple consecutive slices in the z-stack to learn a low dimensional latent space, quantize it and subsequently perform reconstruction using the quantized representation to obtain sharper images. Experiments on a publicly available dataset validate our methodology showing promising results.

eess.IV↗

arXiv2021

Secure, Anonymity-Preserving and Lightweight Mutual Authentication and Key Agreement Protocol for Home Automation IoT Networks

Home automation Internet of Things (IoT) systems have recently become a target for several types of attacks. In this paper, we present an authentication and key agreement protocol for a home automation network based on the ZigBee standard, which connects together a central controller and several end devices. Our scheme performs mutual authentication between end devices and the controller, which is followed by device-to-device communication. The scheme achieves confidentiality, message integrity, anonymity, unlinkability, forward and backward secrecy, and availability. Our scheme uses only simple hash and XOR computations and symmetric key encryption, and hence is resource-efficient. We show using a detailed security analysis and numerical results that our proposed scheme provides better security and anonymity, and is more efficient in terms of computation time, communication cost, and storage cost than schemes proposed in prior works.

cs.CR↗

arXiv2021

Caught in the Act: Core-powered Mass-loss Predictions for Observing Atmospheric Escape

Past studies have demonstrated that atmospheric escape by the core-powered mass-loss mechanism can explain a multitude of observations associated with the radius valley that separates the super-Earth and sub-Neptune planet populations. Complementing such studies, in this work, we present a shortlist of planets that could be losing their atmospheres today if their evolution is indeed primarily dictated by core-powered mass-loss. We use Bayesian inference analysis on our planet evolution and mass-loss model to estimate the posteriors of the parameters that encapsulate the current state of a given planet, given their published masses, radii and host star properties. Our models predict that the following planets could be losing their atmospheres today at a rate $\gtrsim 10^7$ g/s at 50\% confidence level: pi Men c, Kepler-60 d, Kepler-60 b, HD 86226 c, EPIC 249893012 b, Kepler-107 c, HD 219134 b, Kepler-80 e, Kepler-138 d and GJ 9827 d. As a by-product of our Bayesian inference analysis, we were also able to identify planets that most-likely harbor either secondary atmospheres abundant with high mean-molecular weight species, low-density interiors abundant with ices, or both. The planets belonging to this second category are WASP-47 e, Kepler-78 b, Kepler-10 b, CoRoT-7 b, HD 80653 b, 55 Cnc e and Kepler-36 b. While the aforementioned lists are by no means exhaustive, we believe that candidates presented here can serve as useful input for target selection for future surveys and for testing the importance of core-powered mass-loss in individual planetary systems.

astro-ph.EP↗

arXiv2021

Electrical conductivity of strongly magnetized dense quark matter -- possibility of quantum hall effect

We have pointed out the possibility of quantum Hall effect or quantum patterns of transportation in a degenerate strongly magnetized quark matter, which might be expected inside a highly dense compact star. An anisotropic pattern of electrical conductivity and resistivity tensor in classical and quantum cases is explored by considering cyclotron motion and Landau quantization respectively. With increasing magnetic field, classical to quantum transitions are realized through enhanced/reduced resistivity/conductivity with jumping pattern. Considering QCD relaxation time scale of 10 fm, $eB\approx (1-4) m_π^2$ might be considered as strong magnetic field for massless and degenerate quark matter with quark chemical potential $μ\approx 0.2-0.4$ GeV. Beyond these threshold ranges of magnetic field, perpendicular motion of quarks might be stopped and 3 $\rightarrow$ 1 dimensionally reduced conduction picture might be established.

hep-ph↗

arXiv2020

Optimal Sepsis Patient Treatment using Human-in-the-loop Artificial Intelligence

Sepsis is one of the leading causes of death in Intensive Care Units (ICU). The strategy for treating sepsis involves the infusion of intravenous (IV) fluids and administration of antibiotics. Determining the optimal quantity of IV fluids is a challenging problem due to the complexity of a patient's physiology. In this study, we develop a data-driven optimization solution that derives the optimal quantity of IV fluids for individual patients. The proposed method minimizes the probability of severe outcomes by controlling the prescribed quantity of IV fluids and utilizes human-in-the-loop artificial intelligence. We demonstrate the performance of our model on 1122 ICU patients with sepsis diagnosis extracted from the MIMIC-III dataset. The results show that, on average, our model can reduce mortality by 22%. This study has the potential to help physicians synthesize optimal, patient-specific treatment strategies.

cs.AI↗

arXiv2020

ALANET: Adaptive Latent Attention Network forJoint Video Deblurring and Interpolation

Existing works address the problem of generating high frame-rate sharp videos by separately learning the frame deblurring and frame interpolation modules. Most of these approaches have a strong prior assumption that all the input frames are blurry whereas in a real-world setting, the quality of frames varies. Moreover, such approaches are trained to perform either of the two tasks - deblurring or interpolation - in isolation, while many practical situations call for both. Different from these works, we address a more realistic problem of high frame-rate sharp video synthesis with no prior assumption that input is always blurry. We introduce a novel architecture, Adaptive Latent Attention Network (ALANET), which synthesizes sharp high frame-rate videos with no prior knowledge of input frames being blurry or not, thereby performing the task of both deblurring and interpolation. We hypothesize that information from the latent representation of the consecutive frames can be utilized to generate optimized representations for both frame deblurring and frame interpolation. Specifically, we employ combination of self-attention and cross-attention module between consecutive frames in the latent space to generate optimized representation for each frame. The optimized representation learnt using these attention modules help the model to generate and interpolate sharp frames. Extensive experiments on standard datasets demonstrate that our method performs favorably against various state-of-the-art approaches, even though we tackle a much more difficult problem.

cs.CV↗

arXiv2020

Adversarial Knowledge Transfer from Unlabeled Data

While machine learning approaches to visual recognition offer great promise, most of the existing methods rely heavily on the availability of large quantities of labeled training data. However, in the vast majority of real-world settings, manually collecting such large labeled datasets is infeasible due to the cost of labeling data or the paucity of data in a given domain. In this paper, we present a novel Adversarial Knowledge Transfer (AKT) framework for transferring knowledge from internet-scale unlabeled data to improve the performance of a classifier on a given visual recognition task. The proposed adversarial learning framework aligns the feature space of the unlabeled source data with the labeled target data such that the target classifier can be used to predict pseudo labels on the source data. An important novel aspect of our method is that the unlabeled source data can be of different classes from those of the labeled target data, and there is no need to define a separate pretext task, unlike some existing approaches. Extensive experiments well demonstrate that models learned using our approach hold a lot of promise across a variety of visual recognition tasks on multiple standard datasets.

cs.CV↗

arXiv2020

SAFER: Development and Evaluation of an IoT Device Risk Assessment Framework in a Multinational Organization

Users of Internet of Things (IoT) devices are often unaware of their security risks and cannot sufficiently factor security considerations into their device selection. This puts networks, infrastructure and users at risk. We developed and evaluated SAFER, an IoT device risk assessment framework designed to improve users' ability to assess the security of connected devices. We deployed SAFER in a large multinational organization that permits use of private devices. To evaluate the framework, we conducted a mixed-method study with 20 employees. Our findings suggest that SAFER increases users' awareness of security issues. It provides valuable advice and impacts device selection. Based on our findings, we discuss implications for the design of device risk assessment tools, with particular regard to the relationship between risk communication and user perceptions of device complexity.

cs.CR↗

arXiv2020

The evolutionary track of H/He envelope in the observed population of sub-Neptunes and Super-Earths planets

The observational detection of a localized reduction in the small planet occurrence rate, sometimes termed a 'gap', is an exciting discovery because of the implications for planet evolutionary history. This gap appears to define a transition region in which sub-Neptune planets are believed to have lost their H/He envelope, potentially by photoevaporation or core powered mass loss, and have thus been transformed into bare cores terrestrial planets. Here we investigate the transition between sub-Neptunes and super-Earths using a real sample of observed small close-in planets and applying envelope evolution models of the H/He envelope together with the mass-radius diagram and a photoevaporation model. We find that photoevaporation can explain the H/He envelope loss of most super-Earths in 100Myr, although an additional loss mechanism appears necessary in some planets. We explore the possibility that these planets families have different core mass and find a continuum in the primordial population of the strongly irradiated super-Earths and the sub-Neptunes. Our analysis also shows that close-orbiting sub-Neptunes with R < 3.5 R_oplus typically lose ~ 30% of their primordial envelope.

astro-ph.EP↗

arXiv2020

Non-Adversarial Video Synthesis with Learned Priors

Most of the existing works in video synthesis focus on generating videos using adversarial learning. Despite their success, these methods often require input reference frame or fail to generate diverse videos from the given data distribution, with little to no uniformity in the quality of videos that can be generated. Different from these methods, we focus on the problem of generating videos from latent noise vectors, without any reference input frames. To this end, we develop a novel approach that jointly optimizes the input latent space, the weights of a recurrent neural network and a generator through non-adversarial learning. Optimizing for the input latent space along with the network weights allows us to generate videos in a controlled environment, i.e., we can faithfully generate all videos the model has seen during the learning process as well as new unseen videos. Extensive experiments on three challenging and diverse datasets well demonstrate that our approach generates superior quality videos compared to the existing state-of-the-art methods.

eess.IV↗

arXiv2020

Signatures of the Core-Powered Mass-Loss Mechanism in the Exoplanet Population: Dependence on Stellar Properties and Observational Predictions

Recent studies have shown that atmospheric mass-loss powered by the cooling luminosity of a planet's core can explain the observed radius valley separating super-Earths and sub-Neptunes, even without photoevaporation. In this work, we investigate the dependence of this core-powered mass-loss mechanism on stellar mass ($M_\ast$), metallicity ($Z_\ast$) and age ($τ_\ast$). Without making any changes to the underlying planet population, we find that the core-powered mass-loss model yields a shift in the radius valley to larger planet sizes around more massive stars with a slope given by $\text{d log}R_p/\text{d log}M_\ast \simeq 0.35$, in agreement with observations. To first order, this slope is driven by the dependence of core-powered mass-loss on the bolometric luminosity of the host star and is given by $\text{d log}R_p/\text{d log}M_\ast \simeq (3α-2)/36 \simeq 0.33$, where $(L_\ast/L_\odot) = (M_\ast/M_\odot)^α$ is the stellar mass-luminosity relation and $α\simeq 4.6$ for the CKS dataset. We therefore find, in contrast to photoevaporation models, no evidence for a linear correlation between planet and stellar mass, but can't rule it out either. In addition, we show that the location of the radius valley is, to first order, independent of stellar age and metallicity. Since core-powered mass-loss proceeds over Gyr timescales, the abundance of super-Earths relative to sub-Neptunes increases with age but decreases with stellar metallicity. Finally, due the dependence of the envelope's cooling timescale on metallicity, we find that the radii of sub-Neptunes increase with metallicity and decrease with age with slopes given by $\text{d log}R_p/\text{d log}Z_\ast \simeq 0.1$ and $\text{d log}R_p/\text{d log}τ_\ast \simeq -0.1$, respectively. We conclude with a series of observational tests that can differentiate between core-powered mass-loss and photoevaporation models.

astro-ph.EP↗

arXiv2019

Sculpting the Valley in the Radius Distribution of Small Exoplanets as a by-product of Planet Formation: The Core-Powered Mass-Loss Mechanism

Recent observations revealed a bimodal radius distribution of small, short-period exoplanets with a paucity in their occurrence, a radius `valley', around $1.5-2.0$ $R_\oplus$. In this work, we investigate the effect of a planet's own cooling luminosity on its thermal evolution and atmospheric mass-loss (core-powered mass-loss) and determine its observational consequences for the radius distribution of small, close-in exoplanets. Using simple analytical descriptions and numerical simulations, we demonstrate that planetary evolution based on the core-powered mass-loss mechanism alone (i.e., without any photoevaporation) can produce the observed valley in the radius distribution. Our results match the valley's location, shape and slope in planet radius-orbital period parameter space, and the relative magnitudes of the planet occurrence rate above and below the valley. We find that the slope of the valley is, to first order, dictated by the atmospheric mass-loss timescale at the Bondi radius and given by $\text{d log} R_p/ \text{d log} P \simeq 1/(3(1-β))$ which evaluates to $ -0.11$ for $ β\simeq 4$, where $M_c/M_\oplus = (R_c/R_{\oplus})^β (ρ_{c*}/ρ_{\oplus})^{β/3}$ is the mass-radius relation of the core. This choice for $β$ yields good agreement with observations and attests to the significance of internal compression for massive planetary cores. We further find that the location of the valley scales as $ρ_{c*}^{-4/9}$ and that the observed planet population must have predominantly rocky cores with typical water-ice fractions of less than $\sim 20\%$. Furthermore, we show that the relative magnitude of the planet occurrence rate above and below the valley is sensitive to the details of the planet-mass distribution but that the location of the valley is not.

astro-ph.EP↗