Dhirubhai Ambani Institute of Information and Communication Technology

Planning an appropriate sample size for a study involves considering several issues. Two important considerations are cost constraints and variability inherent in the population from which data will be sampled. Methodologists have developed sample size planning methods for two or more populations when testing for equivalence or noninferiority/superiority for a linear contrast of population means. Additionally, cost constraints and variance heterogeneity among populations have also been considered. We extend these methods by developing a theory for sequential procedures for testing the equivalence or noninferiority/superiority for a linear contrast of population means under cost constraints, which we prove to effectively utilize the allocated resources. Our method, due to the sequential framework, does not require prespecified values of unknown population variance(s), something that is historically an impediment to designing studies. Importantly, our method does not require an assumption of a specific type of distribution of the data in the relevant population from which the observations are sampled, as we make our developments in a data distribution-free context. We provide an illustrative example to show how the implementation of the proposed approach can be useful in applied research. (PsycInfo Database Record (c) 2023 APA, all rights reserved).

At nanometer technology nodes, the efficient signal integrity and performance assessment of vast on-chip interconnects are crucial and challenging. For a long time, copper (Cu) has been used as an interconnect material in integrated circuits (ICs). However, as heading towards lower technology nodes, Cu is becoming inadequate to satisfy the requirements for high-speed applications due to its physical limitations. To mitigate this issue, a multiwall carbon nanotube bundle (MWCNTB) is proven to be a better replacement for Cu. Hence, the current work innovatively focuses on modeling, analysis, and performance evaluation of MWCNTB interconnects at 32 nm technology nodes using various machine learning (ML) and neural network (NN) based techniques for signal integrity assessment and fast computation of on-chip interconnect design. Based on the results obtained by comparing the different performance parameters, it is envisaged that NN-based ADAM technique leads to the best-suited model. The developed model is fruitful in evaluating the output performance of the system, such as power-delay-product (PDP), performing parametric analysis, and predicting optimum input design parameters of the driver-interconnect-load (DIL) system. This work utilizes HSPICE and Python electronic design automation tools for its implementation.

We consider a discrete-time single server batch service queue, with a modified bulk service rule, where jobs will process in batches with a minimum batch size of L and maximum batch size K. In addition, we let jobs join the service if the server has started the service and has less than K jobs. The service times of the batches depend on the size of the service batch. We find explicit analytical expressions for the joint and marginal probability distribution of the number of jobs in the system, in the queue, and with the server. We also compute several performance measures and illustrate their dependence on various system parameters. Finally, we match the presented results with some results in the literature and establish a link to the continuous-time counterpart.

ECG is a non-invasive way of determining cardiac health by measuring the electrical activity of the heart. We investigate a novel detection technique for feature points P, QRS and T to diagnose various atrial and ventricular cardiovascular anomalies with ECG signals for ambulatory monitoring. Before the system is worthy of field trials, we validated it with several databases and recorded their response. The QRS complex detection is based on the Pan Tompkins Algorithm and difference operation method that provides positive predictivity, sensitivity and false detection rate of 99.29\%, 99.49\% and 1.29 \% respectively. Proposed novel T wave detection provides sensitivity of 97.78\%. Also, proposed P wave detection provides positive predictivity, sensitivity and false detection rate of 99.43\%, 99.4\% and 1.15\% for the control study (normal subjects) and 82.68\%, 94.3\% and 25.4\% for the case (patients with cardiac anomalies) study respectively. Disease detection such as, arrhythmia is based on standard R-R intervals while myocardial infarction is based on the ST-T deviations where the positive predictivity, sensitivity and accuracy are observed to be 94.6\%, 84.2\% and 85\%, respectively. It should be noted that, since the frontal leads are only used, the anterior myocardial infarction cases are detected with the injury pattern in lead \textit{avl} and ST depression in reciprocal leads. Detection of atrial fibrillation is done for both short and long duration signals using statistical methods using interquartile range and standard deviations, giving very high accuracy, 100\% in most cases. The system hardware for obtaining the 2 lead ECG signal is designed using commercially available off the shelf components. Small field validation of the designed system is performed at a Public Health Centre in Gujarat, India with 42 patients (both cases and controls). We achieved 78.5\% accuracy during the field validation.

The Minimum Dominating Set (MMDS) and Minimum Connected Dominating set (MMCDS) problems are well-studied in the distributed communities due to their numerous applications across the field. These problems are also crucial in wireless ad hoc networks, mainly for the particular type of geometric graphs. We study these problems in geometric graphs such as interval and unit interval graphs. We exploit the underlying geometric structures of these graph classes and present either constant factors distributed algorithms in constant rounds or algorithms with matching lower bounds in the LOCAL communication model.

The Minimum Dominating Set (MDS) and Minimum Connected Dominating set (MCDS) problems are well-studied problems in the distributed computing communities due to their numerous applications across the field. We study these problems in axis-parallel unit square and unit disk graphs. We exploit the underlying geometric structures of these graph classes and present constant round distributed algorithms in the LOCAL communication model. Our results are distributed constant factor approximation algorithms for the MCDS problem in unit square graphs that run in 18 rounds and in unit disk graphs that run in 44 rounds. The message complexity is linear for both the algorithms.

Disease detection and prevention in plants are crucial for generating healthy crops and securing the livelihood of farmers. Leaf wetness duration (LWD), ambient temperature, and relative humidity (RH) are essential parameters that lead to the germination of fungal diseases in plants. In this work, an in-house developed leaf wetness sensor is used to capture LWD, and commercial temperature and humidity sensors are used to record the ambient temperature and humidity, respectively. Subsequently, these sensors are interfaced with an in-house developed internet of things (IoT) enabled electronics and deployed (3 sensor nodes) in the field. We have proposed an attention-based multi-input multi-output neural network (A-MIMONN) to predict diseases using self-collected data. The data for training the model is collected from the three sensors nodes, each comprising of temperature, humidity, and leaf wetness sensors. The designed network is an ensemble of various sub-models, trained individually using data from different sensor nodes. The findings of these individually trained networks are then combined to give the final output. The network is designed to achieve better results by employing the attention mechanism, reinforcing the influence of the most important feature on the predicted results. The average accuracy of the model was found to be about 94%. The model displayed a high average precision of 96% and a high average recall value of 97%. The average F1 score of 97% indicated an excellent balance of precision and accuracy.

Sleep is one of the essential bio-makers for human health. Poor sleep is associated with reduced cognitive performance. With most smartphone users in India being college students, the focus is now on exploring smartphone usage’s impact on students’ sleep. Umpteen news articles in India have reported binge-watching, social media use during the night, and other mobile phone-related interruptions as causes of improper sleep and irregular sleep patterns. However, such studies may involve bias while self-reporting and are limited to a one-time exercise. To understand the reality, we need to accurately quantify the sleep duration, patterns, mobile usage before and after bedtime, number and duration of interruptions. In this first-of-its-kind study in India, we infer novel insights into the sleep patterns of a cohort of 40 college students. We implement a mobile sensing-based approach for the study by installing a custom-developed mobile app on all phones. We extract sleep activity and infer the sleep duration, bed-in and wake-up times, and interruption duration from the sensor data collected from the phone’s built-in sensors. The study brings about new insights into college student sleep patterns and, interestingly, shows that students have a regular sleep cycle and good sleep quality. Only one-fourth of the students revealed irregular sleep patterns, and we did not observe any mobile-related interruptions 30 min past bedtime.

Deep learning models for identification and subsequent mitigation of tokamak plasma disruption have recently shown great promise for reliable predictions for machines other than the one on which it has been trained. The performance of such artificial intelligence (AI)/machine learning (ML) models strongly depends on the training data. Considering the sparse availability of universal high quality data underscores the requirement for synthetic data for the training of the AI/ML models. Synthetic data generation methods reported in the current literature have limitations in terms of quantity, diversity and preserving the temporal dynamics of the experimental seed data (SD). The article presents generative adversarial networks based procedure capable enough to generate unlimited device‐independent temporal evolution of tokamak plasma current. The synthetic data improves with the employment of the classified SD while retaining the characteristics of the original data. The procedure offers a substantial volume of synthetic data with a very impressive diversity, thereby ensuring the requirements for successful AI/ML model training.

Digital data explosion has been driving a demand for robust and reliable data storage medium. The present day digital storage devices presents a big challenge for data scientists to provide reliable, affordable and dense storage medium. In the last decade, natural storage medium such as DNA, bacteria, and protein has been shown as a promising approach for next generation information storage systems. This article is a review that discusses advancements in the natural data storage, which has potential to replace the current physical storage for archival systems and have high storage per unit weight of medium, and is affordable and reliable. Further, the challenges for natural data storage will also be presented. With the international DNA data storage alliance of more than 30 companies and academic alliance in 2020, one would hope that DNA data storage will soon be available commercially in next 5-10 years. The success of the alliance will also motivate researchers to explore the other natural storage medium such as bacteria and protein which are currently in the lab.

This article presents a convolutional neural network (CNN)-based deep-learning (DL) model, inspired from UNet with a series of encoder and decoder units with skip connections, for the simulation of microwave–plasma interaction. The microwave propagation characteristics in complex plasma medium pertaining to transmission, absorption, and reflection primarily depend on the ratio of electromagnetic (EM) wave frequency and electron plasma frequency, and the plasma density profile. The scattering of a plane EM wave with fixed frequency (1 GHz) and amplitude incident on a plasma medium with different Gaussian density profiles (in the range of $1$ $\times$ $10^{17}$ $-$ 1 $\times$ $10^{22}~{\text{m}^{-3}}$ ) have been considered. The training data associated with microwave–plasma interaction has been generated using 2-D finite-difference time-domain (FDTD)-based simulations. The trained DL model is then used to reproduce the scattered electric field values for the 1-GHz incident microwave on different plasma profiles with an error margin of less than 2%. We propose a complete DL-based pipeline to train, validate, and evaluate the model. We compare the results of the network, using various metrics like structural similarity index metric (SSIM) index, average percent error, and mean square error, with the physical data obtained from well-established FDTD-based EM solvers. To the best of our knowledge, this is the first effort toward exploring a DL-based approach for the simulation of complex microwave–plasma interaction. The DL technique proposed in this work is significantly fast when compared to the existing computational techniques and can be used as a new, prospective, and alternative computational approach for investigating microwave–plasma interaction in a real-time scenario.

Twin Extreme Learning Machine models can obtain better generalization ability than the standard Extreme Learning Machine model. But, they require to solve a pair of quadratic programming problems for this. It makes them more complex and computationally expensive than the standard Extreme Learning Machine model. In this paper, we propose two novel time-efficient formulations of the Twin Extreme Learning Machine, which only require the solution of systems of linear equations for obtaining the final classifier. In this sense, they can combine the benefits of the Twin Support Vector Machine and standard Extreme Learning Machine in the true sense. We term our first formulation as ‘Least Squared Twin Extreme Learning Machine’. It minimizes the L2-norm of error variables in its optimization problem. Our second formulation ‘Weighted Linear loss Twin Extreme Learning Machine’ uses the weighted linear loss function for calculating the empirical error, which makes it insensitive towards outliers. Numerical results obtained with multiple benchmark datasets show that proposed formulations are time efficient with better generalization ability. Further, we have used the proposed formulations in the detection of phishing websites and shown that they are much more effective in the detection of phishing websites than other Extreme Learning Machine models.

Monitoring heavy metal pollution in agricultural ecosystems is crucial to ensure environmental safety. Heavy metals interfere with plants’ biochemical characteristics, such as chlorophyll content and photosynthesis, and also influence leaves’ spectral properties. Spectral changes caused by heavy metal stress can easily be measured using proximal sensing or in-field spectroscopy. This research utilizes a combined approach of biochemical and spectral characteristics to evaluate cotton crops’ performance under different heavy metal (Pb & Cd) stress after artificial contamination with the metal under study. A detailed study of spectroscopy and lab-based measurements for chlorophyll and heavy metal content during the crop’s growth cycle revealed some significant findings. Results indicated that the chlorophyll pigments decreased significantly with increased heavy metal levels. Pb accumulation is high in cotton as compared to Cd. The most sensitive stage for the accumulation of Pb is the initial vegetative stage of cotton. The transfer factor from soil to plant was higher for Pb, indicating the feasibility of growing cotton in Pb-contaminated soil. The spectral measurement showed no characteristic changes in standard reflectance spectra due to heavy metal stress. Wavelet decomposition of reflectance spectra amplified the changes indicating Pb stress in cotton during the initial vegetative stage. The significant correlation of greater than − 0.70 between the reconstructed detail wavelet coefficients at the third level of the decomposition in the wavelength range of 651–742 nm suggested that Pb stress caused spectral changes in near-infrared and visible ranges in cotton plants. The effects of Cd stress on the cotton plant were negligible due to less absorption. Thus, detailed wavelet coefficients at the third level of decomposition in the range of 651–742 nm are a potential indicator of Pb stress. The results of this study can provide a basis for quantifying heavy metal stress in a particular region.

For synthetic and voice converted Spoofed Speech Detection (SSD), Instantaneous Frequency (IF)-based features either exploit Hilbert Transform (HT) or Teager-based Energy Separation Algorithm (ESA) to estimate IF. However, HT-based approach leads to poor resolution in time-domain, and ESA-based approach leads to the lack of relative phase information. To that effect, we propose CFCCIF-QESA feature set, which encompasses excellent time resolution as well as relative phase information. Hence, we illustrate the significance of incorporating quadrature-phase component along with in-phase component for SSD of synthetic and voice converted spoof. The proposed feature set is evaluated using various performance metrics, namely, EER, MCC, F-measure, J-statistic, Jaccard Index, and Hamming loss. CFCCIF-QESA achieves a relative decrease in Average EER (AEER) by 22.53% w.r.t. the performance of CFCIF-ESA, for all the attacks in the dataset. Furthermore, results w.r.t. S10 attack (i.e., MaryTTS using unit selection synthesis), which is the most difficult to detect attack, show that CFCCIF-QESA achieves a 29.16% relative decrease in EER as compared to the CFCCIF-ESA. Furthermore, using model-level measures such as, Kullback Leibler Divergence (KLD) and Jenson Shannon Divergence (JSD), we also show the better discriminative ability of our proposed feature set CFCCIF-QESA than the existing CFCCIF-ESA. Finally, the analysis of the latency period for CFCCIF-QESA and CFCCIF-ESA is presented, and it shows better suitability of CFCCIF-QESA w.r.t. deployment in practical SSD systems.

This paper studies and investigates how carbon nano-tube (CNT) bundle serves as future interconnect technology. Interconnects are the main body connecting elements of the integrated circuits (IC). The material that the interconnect is made from is important for its performance. CNTs exhibit excellent properties in terms of electrical, mechanical as well as thermal strength. Its high thermal conductivity and large current carrying capacity makes it an ideal candidate to replace copper interconnects. CNTs can be categorized as single-walled CNT (SWCNT) and Multi-walled CNT (MWCNT). Mixed carbon nanotube bundle has both SWCNTs and MWCNTs. In this paper, different properties of mixed CNT bundle have been analyzed, and effects have been observed by varying the physical properties and dimensions of the bundle. Here, CNT interconnect analyzer (CNIA) tool is used to perform the analysis and simulation. Changes in bundle conductance, bundle inductance, bundle capacitance, bundle propagation delay, drift velocity will be studied by simulating its performance at different values. When bundle width was increased from 2000 to 10,000 nm, bundle conductance increased by almost 433%. Temperature variation was also performed which led us to the conclusion that on increasing temperature by 400 K, bundle conductance increased 8 Ω−1, drift velocity increases by 17.64%, propagation delay decreased by 20%. When bundle length was increased from 20 to 100 µm, inductance and capacitance also increased by 500 and 450%.KeywordsBundle conductanceBundle inductanceBundle capacitanceBundle propagationDelay drift velocity