ISI Digital Commons (Indian Statistical Institute )
Not a member yet
7571 research outputs found
Sort by
LADDER: Revisiting the Cosmic Distance Ladder with Deep Learning Approaches and Exploring Its Applications
No full textWe investigate the prospect of reconstructing the “cosmic distance ladder” of the Universe using a novel deep learning framework called LADDER—Learning Algorithm for Deep Distance Estimation and Reconstruction. LADDER is trained on the apparent magnitude data from the Pantheon Type Ia supernova compilation, incorporating the full covariance information among data points, to produce predictions along with corresponding errors. After employing several validation tests with a number of deep learning models, we pick LADDER as the best-performing one. We then demonstrate applications of our method in the cosmological context, including serving as a model-independent tool for consistency checks for other data sets like baryon acoustic oscillations, calibration of high-redshift data sets such as gamma-ray bursts, and use as a model-independent mock-catalog generator for future probes. Our analysis advocates for careful consideration of machine learning techniques applied to cosmological contexts
Limiting spectral distribution of Toeplitz and Hankel matrices with dependent entries
No full textThis article deals with the limiting spectral distributions (LSD) of symmetric Toeplitz and Hankel matrices with dependent entries. For any fixed integer m≥0, we consider these n×n matrices with entries {Yj(m)/n;j∈Z}, where Yj(m)=∑r=−mmXj+r and {Xk} are i.i.d. random variables with mean zero and variance one. We provide explicit expressions for the LSDs. As a special case (m=0), this article provides an alternate proof for the LSDs of these matrices when the entries are i.i.d. with mean zero and variance one. The method is based on the moment method
Mitigation of extreme events in an excitable system
No full textFormulating mitigation strategies is one of the main aspect in the dynamical study of extreme events. Apart from the effective control, easy implementation of the devised tool should also be given importance. In this work, we analyze the mitigation of extreme events in a coupled FitzHugh–Nagumo (FHN) neuron model utilizing an easily implementable constant bias analogous to a constant DC stimulant. We report the route through which the extreme events gets mitigated in Two, Three and N-coupled FHN systems. In all the three cases, extreme events in the observable x¯ gets suppressed. We confirm our results with the probability distribution function of peaks, dmax plot and probability plots. Here dmax is a measure of number of standard deviations that crosses the average amplitude corresponding to x¯max. Interestingly, we found that constant bias suppresses the extreme events without changing the collective frequency of the system
Multi-lingual handwriting recovery framework based on convolutional denoising autoencoder with attention model
No full textFor several decades, no satisfactory solutions have been provided to the problem of offline handwriting recognition. In the field of online recognition, researchers have had more successful performance, but the ability to extract dynamic information from static images has not been well explored yet. In this paper, we introduce a novel multi-lingual word handwriting recovery framework based on a convolutional denoising autoencoder with an attention model for pen up/down, velocity and temporal order recovery. The proposed framework consists of extracting robust features from a handwriting image using a stacked denoising autoencoder and an encoder Bidirectional Gated Recurrent Unit (BGRU) model. Then, the obtained vectors are decoded to produce an online script with dynamic characteristics using a BGRU with temporal attention. Evaluation is done on a Latin and Arabic Online and offline handwriting character / word databases and the proposed framework achieves high competitive results
On some statistical properties of a stationary Gaussian process in the presence of measurement errors
No full textProcess outputs of many production processes like chemical, food processing and pharmaceutical industry follow a stationary Gaussian process. Some amount of measurement error always present in the measured data due to inaccurate measuring processes. Throughout this paper, we discuss some statistical properties like the mean and variance of a stationary Gaussian process when observed data are affected by measurement errors. As a special case, we discuss a stationary autoregressive process of order one with Gaussian white noise where measurement error follows an independent Gaussian distribution
Optimum selection of nature-inspired texture pattern for cutting tool surface using an integrated multi-criteria decision-making approach: a comparative analysis
No full textPurpose: This paper aims to focus on the selection of an appropriate nature-inspired texture pattern for cutting tool tribological surface. The selection process uses the recognized skin textures of different snakes scrolling on highly rough and projected surface conditions to analyze suitability of texture based on the texture geometry and machining conditions. The work also aims to propose a texture pattern selection process to incorporate on cutting tool tribological surface. Design/methodology/approach: The selection of alternative nature-inspired texture patterns based on the texture pattern geometry and machining properties leads to a multi-criteria decision-making problem. Thirteen criteria are considered for selecting an appropriate texture pattern among 14 alternatives, i.e. nature-inspired texture patterns. In the present work, an integrated analytical hierarchy process (AHP)-TOPSIS, AHP-multi-objective optimization on the basis of ratio analysis (MOORA) and AHP-Vlse Kriterijumska Optimizacija Kompromisno Resenje (VIKOR) approaches have been proposed for the selection of an appropriate nature-inspired texture pattern. AHP is used for the formulation of decision-making matrix and criteria weight calculations and ranking of alternatives is done by three methods. Spearman’s correlation compared and found positive relations between rank assigned by methods. Experimental validation is done in Lathe for selected texture effects. Findings: The texture parameters C-1 (Width of texture) and C-2 (Depth of texture) are found significant, while T-2 (Blended Krait) and T-6 (Banded Racer-1) texture is found optimal to generate on cutting tool surface. Research limitations/implications: Only some nature-inspired texture patterns have been recognized before the selection; an infinite number of textures are available in nature. The size of the texture pattern is difficult to identify by the selection process because each texture pattern may have different effects on tribological surfaces. Practical implications: The proposed selection methodology of nature-inspired texture patterns will help identify optimal texture geometry for specific tribological applications. The nature-inspired texture patterned tool has a significant impact on the cutting force and temperature due to its tribological effect on the cutting tool surface; it decreases the power required for machining. The machining characteristics like roughness are found to decrease by using nature-inspired texture patterned tools. Social implications: Various nature-inspire texture studies to generate specific effects on the tribological surfaces may be started study for the surface of aircraft, ships, bearings, etc. Small and big fabrication industries may benefit by decreasing the cost of machining using nature-inspired texture-patterned tools. Research society will pay attention to nature’s inspiration. Originality/value: Novel snake-skin-inspired texture patterns are recognized and hybrid MCDM methods are proposed to select optimal texture pattern. Proposed method used single time normalization to effectively rank the alternatives. The insights gained from this research can be extrapolated to address similar challenges in selecting nature-inspired textures for various applications. Peer review: The peer review history for this article is available at: https://publons.com/publon/10.1108/ILT-05-2024-0163
Efficient Reduction of Resources and Noise in Discrete Quantum Computing Circuits
No full textQuantum computers have shifted from a subject of theoretical interest to reality in recent years with multiple devices now available at research industry labs such as IBM, Google, and IonQ. However, quantum systems are highly susceptible to noise. Interaction with the environment corrupts the information content, leading to unreliable computation. Near-term quantum computers do not have sufficient qubits to incorporate error correction. Therefore, other mechanisms are studied to lower the effect of noise. Quantum Approximate Optimization Algorithm (QAOA) is an algorithm family for finding approximate solutions to combinatorial optimization problems. Any such problem can be represented as a graph, and the number of 2-qubit gates in the corresponding circuit scales linearly with the number of edges. 2-qubit gates are one of the noisiest components in current hardware. This thesis proposes three hardware-independent algorithms to lower the number of 2-qubit gates while ensuring functional equivalence. Finally, a modification to these algorithms is proposed when the underlying hardware connectivity and layout of the circuit is known. Another method, called circuit cutting, where a circuit is partitioned into multiple smaller subcircuits, is shown to effectively lower noise. The subcircuits are computed individually, and the outcome is constructed on a classical computer. This thesis proposes two error mitigation techniques, targeted particularly for circuit cutting, which significantly improve the fidelity of the outcome. These techniques for lowering the effect of noise are not sufficient for arbitrary long quantum computation; there quantum error correction (QECC) is mandated. The second part of this thesis shows the challenges of designing a ternary QECC from its binary counterpart. Naive efforts require two-step error correction, leading to a significant increase in the gate cost of the QECC circuit. Next, a necessary condition for stabilizer formulation is provided which allows easy carry-over of binary QECC to ternary, making error correction a single step. For near-term devices, the decomposition of a 3-qubit Toffoli gate by temporary access to higher dimensions is shown to provide an exponential reduction in the depth of the decomposed circuit. The final chapter studies provides an analytical criterion for which the resource requirement of qutrit-assisted decomposition, when equipped with error correction, remains lower than the qubit-only decomposition. Overall, this thesis contributes both to near-term and long-term quantum computing. The findings from Part I provide useful methods to improve the performance of algorithms in current quantum devices. Part II gives valuable insight into the challenges of incorporating near-term methods in conjunction with error correction and the design of higher dimensional QECC from binary codes
Some stochastic comparison results for frailty and resilience models
No full textAbstract.: Frailty and resilience models provide a way to introduce random effects in hazard and reversed hazard rate modeling by random variables, called frailty and resilience random variables, respectively, to account for unobserved or unexplained heterogeneity among experimental units. This article investigates the effects of frailty and resilience random variables on the baseline random variables using some shifted stochastic orders based on some ageing properties of the baseline random variables. Relevant examples are presented to illustrate the findings, and the applicability of the results is demonstrated with real-world data, highlighting their relevance in practical scenarios
Stability analysis of synchronization in long-range temporal networks using theory of dichotomy
No full textMost of the previous studies on the stability analysis of synchronization in static or time-varying networks are based on the master stability function approach, which is a semi-analytical concept. The necessary and sufficient conditions for synchronization in time-varying networks are challenging problems since the last few years. We focus on the stability analysis of synchronization in time-varying networks, particularly long-range networks. The use of dichotomy theory to derive sufficient conditions for synchronization in this context is an interesting approach. The incorporation of long-range interactions adds complexity and might lead to larger regions of synchronization, providing valuable insights into the dynamics of such networks. Analyzing the co-action of the time-varying nature in the network topology and long-range interactions is a relevant and challenging task, especially when the network is not synchronized. This work appears to explore the interplay between these factors and their impact on synchronization. Additionally, the numerical study considering long-range connections governed by a power-law within the framework of an Erdös-Rényi random network is a practical way to validate and test the analytical results. It is good to see that we are exploring the effects of varying parameters such as rewiring probability, coupling strength, and power-law exponent on the synchronization state
Stochastic comparison of the second-order statistics arising from exponentiated location-scale model
No full textIn this article, we consider stochastic comparisons between second-order statistics arising from general exponentiated location-scale models. When the random variables are independent, we establish usual stochastic and hazard rate orders between second-order statistics. Further, similar ordering results are obtained when the random observations are dependent. Some applications of the established results are presented