arXiv subjects

Abhishek Srivastava

Publications and source records attributed to Abhishek Srivastava.

At least 37 records · Page 2Linked to original sources

arXiv2021

AGA-GAN: Attribute Guided Attention Generative Adversarial Network with U-Net for Face Hallucination

The performance of facial super-resolution methods relies on their ability to recover facial structures and salient features effectively. Even though the convolutional neural network and generative adversarial network-based methods deliver impressive performances on face hallucination tasks, the ability to use attributes associated with the low-resolution images to improve performance is unsatisfactory. In this paper, we propose an Attribute Guided Attention Generative Adversarial Network which employs novel attribute guided attention (AGA) modules to identify and focus the generation process on various facial features in the image. Stacking multiple AGA modules enables the recovery of both high and low-level facial structures. We design the discriminator to learn discriminative features exploiting the relationship between the high-resolution image and their corresponding facial attribute annotations. We then explore the use of U-Net based architecture to refine existing predictions and synthesize further facial details. Extensive experiments across several metrics show that our AGA-GAN and AGA-GAN+U-Net framework outperforms several other cutting-edge face hallucination state-of-the-art methods. We also demonstrate the viability of our method when every attribute descriptor is not known and thus, establishing its application in real-world scenarios.

cs.CV↗

arXiv2021

Exploiting Multi-Scale Fusion, Spatial Attention and Patch Interaction Techniques for Text-Independent Writer Identification

Text independent writer identification is a challenging problem that differentiates between different handwriting styles to decide the author of the handwritten text. Earlier writer identification relied on handcrafted features to reveal pieces of differences between writers. Recent work with the advent of convolutional neural network, deep learning-based methods have evolved. In this paper, three different deep learning techniques - spatial attention mechanism, multi-scale feature fusion and patch-based CNN were proposed to effectively capture the difference between each writer's handwriting. Our methods are based on the hypothesis that handwritten text images have specific spatial regions which are more unique to a writer's style, multi-scale features propagate characteristic features with respect to individual writers and patch-based features give more general and robust representations that helps to discriminate handwriting from different writers. The proposed methods outperforms various state-of-the-art methodologies on word-level and page-level writer identification methods on three publicly available datasets - CVL, Firemaker, CERUG-EN datasets and give comparable performance on the IAM dataset.

cs.CV↗

arXiv2021

GMSRF-Net: An improved generalizability with global multi-scale residual fusion network for polyp segmentation

Colonoscopy is a gold standard procedure but is highly operator-dependent. Efforts have been made to automate the detection and segmentation of polyps, a precancerous precursor, to effectively minimize missed rate. Widely used computer-aided polyp segmentation systems actuated by encoder-decoder have achieved high performance in terms of accuracy. However, polyp segmentation datasets collected from varied centers can follow different imaging protocols leading to difference in data distribution. As a result, most methods suffer from performance drop and require re-training for each specific dataset. We address this generalizability issue by proposing a global multi-scale residual fusion network (GMSRF-Net). Our proposed network maintains high-resolution representations while performing multi-scale fusion operations for all resolution scales. To further leverage scale information, we design cross multi-scale attention (CMSA) and multi-scale feature selection (MSFS) modules within the GMSRF-Net. The repeated fusion operations gated by CMSA and MSFS demonstrate improved generalizability of the network. Experiments conducted on two different polyp segmentation datasets show that our proposed GMSRF-Net outperforms the previous top-performing state-of-the-art method by 8.34% and 10.31% on unseen CVC-ClinicDB and unseen Kvasir-SEG, in terms of dice coefficient.

eess.IV↗

arXiv2021

PAANet: Progressive Alternating Attention for Automatic Medical Image Segmentation

Medical image segmentation can provide detailed information for clinical analysis which can be useful for scenarios where the detailed location of a finding is important. Knowing the location of disease can play a vital role in treatment and decision-making. Convolutional neural network (CNN) based encoder-decoder techniques have advanced the performance of automated medical image segmentation systems. Several such CNN-based methodologies utilize techniques such as spatial- and channel-wise attention to enhance performance. Another technique that has drawn attention in recent years is residual dense blocks (RDBs). The successive convolutional layers in densely connected blocks are capable of extracting diverse features with varied receptive fields and thus, enhancing performance. However, consecutive stacked convolutional operators may not necessarily generate features that facilitate the identification of the target structures. In this paper, we propose a progressive alternating attention network (PAANet). We develop progressive alternating attention dense (PAAD) blocks, which construct a guiding attention map (GAM) after every convolutional layer in the dense blocks using features from all scales. The GAM allows the following layers in the dense blocks to focus on the spatial locations relevant to the target region. Every alternate PAAD block inverts the GAM to generate a reverse attention map which guides ensuing layers to extract boundary and edge-related information, refining the segmentation process. Our experiments on three different biomedical image segmentation datasets exhibit that our PAANet achieves favourable performance when compared to other state-of-the-art methods.

eess.IV↗

arXiv2021

BEIR: A Heterogenous Benchmark for Zero-shot Evaluation of Information Retrieval Models

Existing neural information retrieval (IR) models have often been studied in homogeneous and narrow settings, which has considerably limited insights into their out-of-distribution (OOD) generalization capabilities. To address this, and to facilitate researchers to broadly evaluate the effectiveness of their models, we introduce Benchmarking-IR (BEIR), a robust and heterogeneous evaluation benchmark for information retrieval. We leverage a careful selection of 18 publicly available datasets from diverse text retrieval tasks and domains and evaluate 10 state-of-the-art retrieval systems including lexical, sparse, dense, late-interaction and re-ranking architectures on the BEIR benchmark. Our results show BM25 is a robust baseline and re-ranking and late-interaction-based models on average achieve the best zero-shot performances, however, at high computational costs. In contrast, dense and sparse-retrieval models are computationally more efficient but often underperform other approaches, highlighting the considerable room for improvement in their generalization capabilities. We hope this framework allows us to better evaluate and understand existing retrieval systems, and contributes to accelerating progress towards better robust and generalizable systems in the future. BEIR is publicly available at https://github.com/UKPLab/beir.

cs.IR↗

arXiv2021

An Efficient Anomaly Detection Approach using Cube Sampling with Streaming Data

Anomaly detection is critical in various fields, including intrusion detection, health monitoring, fault diagnosis, and sensor network event detection. The isolation forest (or iForest) approach is a well-known technique for detecting anomalies. It is, however, ineffective when dealing with dynamic streaming data, which is becoming increasingly prevalent in a wide variety of application areas these days. In this work, we extend our previous work by proposed an efficient iForest based approach for anomaly detection using cube sampling that is effective on streaming data. Cube sampling is used in the initial stage to choose nearly balanced samples, significantly reducing storage requirements while preserving efficiency. Following that, the streaming nature of data is addressed by a sliding window technique that generates consecutive chunks of data for systematic processing. The novelty of this paper is in applying Cube sampling in iForest and calculating inclusion probability. The proposed approach is equally successful at detecting anomalies as existing state-of-the-art approaches, requiring significantly less storage and time complexity. We undertake empirical evaluations of the proposed approach using standard datasets and demonstrate that it outperforms traditional approaches in terms of Area Under the ROC Curve (AUC-ROC) and can handle high-dimensional streaming data.

cs.LG↗

arXiv2021

An Energy Efficient Health Monitoring Approach with Wireless Body Area Networks

Wireless Body Area Networks (WBANs) comprise a network of sensors subcutaneously implanted or placed near the body surface and facilitate continuous monitoring of health parameters of a patient. Research endeavours involving WBAN are directed towards effective transmission of detected parameters to a Local Processing Unit (LPU, usually a mobile device) and analysis of the parameters at the LPU or a back-end cloud. An important concern in WBAN is the lightweight nature of WBAN nodes and the need to conserve their energy. This is especially true for subcutaneously implanted nodes that cannot be recharged or regularly replaced. Work in energy conservation is mostly aimed at optimising the routing of signals to minimise energy expended. In this paper, a simple yet innovative approach to energy conservation and detection of alarming health status is proposed. Energy conservation is ensured through a two-tier approach wherein the first tier eliminates `uninteresting' health parameter readings at the site of a sensing node and prevents these from being transmitted across the WBAN to the LPU. A reading is categorised as uninteresting if it deviates very slightly from its immediately preceding reading and does not provide new insight on the patient's well being. In addition to this, readings that are faulty and emanate from possible sensor malfunctions are also eliminated. These eliminations are done at the site of the sensor using algorithms that are light enough to effectively function in the extremely resource-constrained environments of the sensor nodes. We notice, through experiments, that this eliminates and thus reduces around 90% of the readings that need to be transmitted to the LPU leading to significant energy savings. Furthermore, the proper functioning of these algorithms in such constrained environments is confirmed and validated over a hardware simulation set up. The second tier of assessment includes a proposed anomaly detection model at the LPU that is capable of identifying anomalies from streaming health parameter readings and indicates an adverse medical condition. In addition to being able to handle streaming data, the model works within the resource-constrained environments of an LPU and eliminates the need of transmitting the data to a back-end cloud, ensuring further energy savings. The anomaly detection capability of the model is validated using data available from the critical care units of hospitals and is shown to be superior to other anomaly detection techniques.

eess.SP↗

arXiv2021

A mmWave Oscillator Design Utilizing High-Q Active-Mode On-Chip MEMS Resonators for Improved Fundamental Limits of Phase Noise

(RFT) allows very high-Q active mode resonators, promising crystal-less monolithic clock generation for mmWave systems. However, there is a strong need for design of mmWave oscillators that utilize the high-Q of active-mode RFT (AM-RFT) optimally, while handling unique challenges such as resonator's low electromechanical transduction. In this brief, we develop a theory and through design and post-layout simulations in 14 nm Global Foundry process, we show the first active oscillator with AM-RFT at 30 GHz, which improves the fundamental limits of phase noise and figure-of-merit as compared to the oscillators with conventional LC resonators. For AM-RFT with Q factor of 10K, post layout simulation results show that the proposed oscillator exhibits phase noise less than -140 dBc per Hz and figure-of-merit greater than 228 dBc per Hz at 1 MHz offset for 30 GHz center frequency, which are more than 25 dB better than the existing monolithic LC oscillators.

eess.SP↗

arXiv2021

Conditional LSTM-GAN for Melody Generation from Lyrics

Melody generation from lyrics has been a challenging research issue in the field of artificial intelligence and music, which enables to learn and discover latent relationship between interesting lyrics and accompanying melody. Unfortunately, the limited availability of paired lyrics-melody dataset with alignment information has hindered the research progress. To address this problem, we create a large dataset consisting of 12,197 MIDI songs each with paired lyrics and melody alignment through leveraging different music sources where alignment relationship between syllables and music attributes is extracted. Most importantly, we propose a novel deep generative model, conditional Long Short-Term Memory - Generative Adversarial Network (LSTM-GAN) for melody generation from lyrics, which contains a deep LSTM generator and a deep LSTM discriminator both conditioned on lyrics. In particular, lyrics-conditioned melody and alignment relationship between syllables of given lyrics and notes of predicted melody are generated simultaneously. Experimental results have proved the effectiveness of our proposed lyrics-to-melody generative model, where plausible and tuneful sequences can be inferred from lyrics.

cs.AI↗

arXiv2020

Conditional Hybrid GAN for Sequence Generation

Conditional sequence generation aims to instruct the generation procedure by conditioning the model with additional context information, which is a self-supervised learning issue (a form of unsupervised learning with supervision information from data itself). Unfortunately, the current state-of-the-art generative models have limitations in sequence generation with multiple attributes. In this paper, we propose a novel conditional hybrid GAN (C-Hybrid-GAN) to solve this issue. Discrete sequence with triplet attributes are separately generated when conditioned on the same context. Most importantly, relational reasoning technique is exploited to model not only the dependency inside each sequence of the attribute during the training of the generator but also the consistency among the sequences of attributes during the training of the discriminator. To avoid the non-differentiability problem in GANs encountered during discrete data generation, we exploit the Gumbel-Softmax technique to approximate the distribution of discrete-valued sequences.Through evaluating the task of generating melody (associated with note, duration, and rest) from lyrics, we demonstrate that the proposed C-Hybrid-GAN outperforms the existing methods in context-conditioned discrete-valued sequence generation.

cs.AI↗

arXiv2020

Magnetic Damping in Epitaxial Fe Alloyed with Vanadium and Aluminum

To develop low-moment, low-damping metallic ferromagnets for power-efficient spintronic devices, it is crucial to understand how magnetic relaxation is impacted by the addition of nonmagnetic elements. Here, we compare magnetic relaxation in epitaxial Fe films alloyed with light nonmagnetic elements of V and Al. FeV alloys exhibit lower intrinsic damping compared to pure Fe, reduced by nearly a factor of 2, whereas damping in FeAl alloys increases with Al content. Our experimental and computational results indicate that reducing the density of states at the Fermi level, rather than the average atomic number, has a more significant impact in lowering damping in Fe alloyed with light elements. Moreover, FeV is confirmed to exhibit an intrinsic Gilbert damping parameter of $\simeq$0.001, among the lowest ever reported for ferromagnetic metals.

cond-mat.mtrl-sci↗

arXiv2020

Implementing Stepped Pooled Testing for Rapid COVID-19 Detection

COVID-19, a viral respiratory pandemic, has rapidly spread throughout the globe. Large scale and rapid testing of the population is required to contain the disease, but such testing is prohibitive in terms of resources, cost and time. Recently RT-PCR based pooled testing has emerged as a promising way to boost testing efficiency. We introduce a stepped pooled testing strategy, a probability driven approach which significantly reduces the number of tests required to identify infected individuals in a large population. Our comprehensive methodology incorporates the effect of false negative and positive rates to accurately determine not only the efficiency of pooling but also it's accuracy. Under various plausible scenarios, we show that this approach significantly reduces the cost of testing and also reduces the effective false positive rate of tests when compared to a strategy of testing every individual of a population. We also outline an optimization strategy to obtain the pool size that maximizes the efficiency of pooling given the diagnostic protocol parameters and local infection conditions.

q-bio.PE↗

arXiv2020

Analysis and Design Considerations for Achieving the Fundamental Limits of Phase Noise in mmWave Oscillators with On-Chip MEMS Resonator

Very small electromechanical coupling coefficient in micro-electromechanical systems (MEMS) or acoustic resonators is quite of a concern for oscillator performance, specially at mmWave frequencies. This small coefficient is the manifestation of the small ratio of motional capacitance to static capacitance in the resonators. This work provides a general solution to overcome the problem of relatively high static capacitance at mmWave frequencies and presents analysis and design techniques for achieving extremely low phase noise and a very high figure-of-merit (FoM) in an on-chip MEMS resonator based mmWave oscillator. The proposed analysis and techniques are validated with design and simulation of a 30 GHz oscillator with MEMS resonator having quality factor of 10,000 in 14 nm GF technology. Post layout simulation results show that it achieves a phase noise of -132 dBc/Hz and FoM of 217 dBc/Hz at offset of 1 MHz.

eess.SP↗

arXiv2020

Sub-Nanosecond Spin-Transfer Torque in an Ensemble of Superparamagnetic-Like Nanomagnets

Spin currents can exert spin-transfer torques on magnetic systems even in the limit of vanishingly small net magnetization, as is the case for antiferromagnets. Here, we experimentally show that a spin-transfer torque is operative in a material with weak, short-range magnetic order -- namely, a macroscopic ensemble of superparamagnetic-like Co nanomagnets. We employ element- and time-resolved X-ray ferromagnetic resonance (XFMR) spectroscopy to directly detect sub-ns dynamics of the Co nanomagnets, excited into precession with cone angle $\geq$0.003$^{\circ}$ by an oscillating spin current. XFMR measurements reveal that as the net moment of the ensemble decreases, the strength of the spin-transfer torque increases relative to those of magnetic field torques. Our findings point to spin-transfer torque as an effective way to manipulate the state of nanomagnet ensembles at sub-ns timescales.

cond-mat.mes-hall↗

arXiv2020

Conductivity-Like Gilbert Damping due to Intraband Scattering in Epitaxial Iron

Confirming the origin of Gilbert damping by experiment has remained a challenge for many decades, even for simple ferromagnetic metals. In this Letter, we experimentally identify Gilbert damping that increases with decreasing electronic scattering in epitaxial thin films of pure Fe. This observation of conductivity-like damping, which cannot be accounted for by classical eddy current loss, is in excellent quantitative agreement with theoretical predictions of Gilbert damping due to intraband scattering. Our results resolve the longstanding question about a fundamental damping mechanism and offer hints for engineering low-loss magnetic metals for cryogenic spintronics and quantum devices.

cond-mat.mtrl-sci↗

arXiv2019

HTS: A Hardware Task Scheduler for Heterogeneous Systems

As the Moore's scaling era comes to an end, application specific hardware accelerators appear as an attractive way to improve the performance and power efficiency of our computing systems. A massively heterogeneous system with a large number of hardware accelerators along with multiple general purpose CPUs is a promising direction, but pose several challenges in terms of the run-time scheduling of tasks on the accelerators and design granularity of accelerators. This paper addresses these challenges by developing an example heterogeneous system to enable multiple applications to share the available accelerators. We propose to design accelerators at a lower abstraction to enable applications to be broken down into tasks that can be mapped on several accelerators. We observe that several real-life workloads can be broken down into common primitives that are shared across many workloads. Finally, we propose and design a hardware task scheduler inspired by the hardware schedulers in out-of-order superscalar processors to efficiently utilize the accelerators in the system by scheduling tasks in out-of-order and even speculatively. We evaluate the proposed system on both real-life and synthetic benchmarks based on Digital Signal Processing~(DSP) applications. Compared to executing the benchmark on a system with sequential scheduling, proposed scheduler achieves up to 12x improvement in performance.

cs.OS↗

arXiv2019

Predicting Onset of Dementia in Parkinson's Disease Patients

Alzheimer's disease (AD) and Parkinson's disease (PD) are the two most common neurodegenerative disorders in humans. Because a significant percentage of patients have clinical and pathological features of both diseases, it has been hypothesized that the patho-cascades of the two diseases overlap. Despite this evidence, these two diseases are rarely studied in a joint manner. In this paper, we utilize clinical, imaging, genetic, and biospecimen features to cluster AD and PD patients into the same feature space. By training a machine learning classifier on the combined feature space, we predict the disease stage of patients two years after their baseline visits. We observed a considerable improvement in the prediction accuracy of Parkinson's dementia patients due to combined training on Alzheimer's and Parkinson's patients, thereby affirming the claim that these two diseases can be jointly studied.

q-bio.QM↗

arXiv2019

Vectorial observation of the spin Seebeck effect in epitaxial NiFe$_2$O$_4$ thin films with various magnetic anisotropy contributions

We have developed a vectorial type of measurement for the spin Seebeck effect (SSE) in epitaxial NiFe$_2$O$_4$ thin films which have been grown by pulsed laser deposition on MgGa$_2$O$_4$ (MGO) with (001) and (011) orientation as well as CoGa$_2$O$_4$ (011) (CGO), thus varying the lattice mismatch and crystal orientation. We confirm that a large lattice mismatch leads to strain anisotropy in addition to the magnetocrystalline anisotropy in the thin films using vibrating sample magnetometry and ferromagnetic resonance measurements. Moreover, we show that the existence of a magnetic strain anisotropy in NiFe$_2$O$_4$ thin films significantly impacts the shape and magnitude of the magnetic-field-dependent SSE voltage loops. We further demonstrate that bidirectional field-dependent SSE voltage curves can be utilized to reveal the complete magnetization reversal process, which establishes a vectorial magnetometry technique based on a spin caloric effect.

cond-mat.mes-hall↗