ISI Digital Commons (Indian Statistical Institute )
Not a member yet
7571 research outputs found
Sort by
Understanding the prospects of human-wildlife coexistence: a conceptual framework
No full textHuman-wildlife interactions can range from reverence to extreme conflict. Conservationists have come to the realization that humans and wildlife have always coexisted together in shared landscapes across the globe. Thus, understanding and acting upon the prospects of human-wildlife coexistence (HWCo) is now a crucial component of biodiversity conservation to sustain it. HWCo is a state where humans and wildlife share spaces by exposing each other to tolerable levels of risks and disadvantages. HWCo transpires as a result of interplay between a number of perceived and behavioral factors, some of which are interdependent on one another. Through this framework, we find ways to identify these factors, which can then be used to evaluate HWCo and understand the drivers of HWCo. Therefore, the current article focuses on changing this paradigm in HWCo research. We suggest three continuums involving three crucial factors viz., space-use by wildlife, daily activity pattern of wildlife, and human attitude towards wildlife, be used to obtain a cumulative value signifying HWCo for a particular species/taxon in a shared landscape. We propose that these factors be measured simultaneously on a predefined scale, which will allow it to become relative, and will further allow cross-site comparisons. This preliminary framework is expected to enable scientists and researchers to visualize the complexity and dynamicity embedded within human-wildlife interactions through modeling. The evaluation on a continuum is especially effective when positive or negative interactions between humans and wildlife are not obvious
Geometric Covering Number: Covering Points with Curves
No full textGiven a point set, mostly a grid in our case, we seek upper and lower bounds on the number of curves that are needed to cover the point set. We say a curve covers a point if the curve passes through the point. We consider such coverings by monotonic curves, lines, orthoconvex curves, circles, etc. We also study a problem that is converse of the covering problem – if a set of n2 points in the plane is covered by n lines then can we say something about the configuration of the points
GraphKD: Exploring Knowledge Distillation Towards Document Object Detection with Structured Graph Creation
No full textObject detection in documents is a key step to automate the structural elements identification process in a digital or scanned document through understanding the hierarchical structure and relationships between different elements. Large and complex models, while achieving high accuracy, can be computationally expensive and memory-intensive, making them impractical for deployment on resource constrained devices. Knowledge distillation allows us to create small and more efficient models that retain much of the performance of their larger counterparts. Here we present a graph-based knowledge distillation framework to correctly identify and localize the document objects in a document image. Here, we design a structured graph with nodes containing proposal-level features and edges representing the relationship between the different proposal regions. Also, to reduce text bias an adaptive node sampling strategy is designed to prune the weight distribution and put more weightage on non-text nodes. We encode the complete graph as a knowledge representation and transfer it from the teacher to the student through the proposed distillation loss by effectively capturing both local and global information concurrently. Extensive experimentation on competitive benchmarks demonstrates that the proposed framework outperforms the current state-of-the-art approaches. The code will be available at: github.com/ayanban011/GraphKD
Geometric covering via extraction Theorem
No full textIn this work, we address the following question. Suppose we are given a set D of positive-weighted disks and a set T of n points in the plane, such that each point of T is contained in at least two disks of D. Then is there always a subset S of D such that the union of the disks in S contains all the points of T and the total weight of the disks of D that are not in S is at least a constant fraction of the total weight of the disks in D? In our work, we prove the Extraction Theorem that answers this question in the affirmative. Our constructive proof heavily exploits the geometry of disks, and in the process, we make interesting connections between our work and the literature on local search for geometric optimization problems. The Extraction Theorem helps to design the first polynomial-Time O(1)-Approximations for two important geometric covering problems involving disks
User-Authenticated Device-Independent Quantum Secure Direct Communication Protocol
No full textDevice-Independent Quantum Secure Direct Communication (DI-QSDC) enhances quantum cryptography by enabling secure message transmission without relying on the trustworthiness of the devices involved. This approach mitigates risks associated with compromised or untrusted devices, common in traditional quantum communication. In this paper, we propose the first of its kind DI-QSDC protocol with user identity authentication. This ensures the authenticity of both the sender and receiver prior to message exchange. We then discuss the security of the proposed protocol against common attacks, demonstrating that no eavesdropper gains any information from either the quantum or the classical channel. Next, we implement the protocol on IBM\u27s quantum hardware and evaluate its performance in a realistic noisy environment. Additionally, by simulating common attack models, we showcase that the protocol is secure against any eavesdropper in the channel. These findings highlight the protocol\u27s robust security and practical feasibility for real-world secure quantum communication
Kinetic exchange models of income and wealth distribution: Self organization and poverty level
No full textIn this book chapter, we draw the reader to a brief review of the different Kinetic Exchange Models that have gradually developed for markets and how they can be employed to quantitatively study inequalities (the Gini Index and the Kolkata Index) that pervade real economies. Since a many-body economical market can be studied using tested laws of physics, these models have the freedom to incorporate provisions like inclusion of individual saving behaviors while trading and rendering these saving behaviors to be time-independent and time-dependent respectively. This is to observe when and how the well known exponential distribution for no-saving gives rise to a distribution with a most probable income. A review of the earlier cases along with their implementation with a bias, where the population in the low-income bracket of the society is favored by selecting one of the traders in the trading process to be the poorest of all, follows. The biased selection of agents ultimately giving rise to self-organizing features in the money distribution has also been reviewed in detail, using the behavior of the Self-Organized Poverty Level. In the end, we elucidate the recent endeavors of the KEMs in finding answers to the growing condensation of wealth in the hands of a countable few rich people and providing probable solutions that can curb further expansion of oligarchic societies
Spectrum of High-Dimensional Sample Covariance and Related Matrices: A Selective Review
No full textThis is a selective review on the behavior of the high-dimensional sample covariance matrix, S=n-1XX∗, the most important random matrix in high-dimensional statistics, and some related matrices. The asymptotic distribution of the empirical spectrum of (centered and scaled) S is either the semi-circular law or the Marčenko–Pastur law, depending on how the dimensions n and p grow. From a non-commutative probability point of view, independent copies of S converge algebraically to freely independent semi-circular or compound free Poisson variables. We also discuss the asymptotic normality of the linear spectral statistics, and the convergence of the maximum eigenvalue to a Tracy–Widom law. Some related matrices are also discussed. These include the separable, generalized, and cross-covariance matrices, Kendall’s τ and Spearman’s ρ matrices, and the autocovariance matrices in one and high dimensions. Finally, we present some statistical applications
Undernutrition Among Infants Through Composite Index of Anthropometric Failures
No full textObjectives of the Study: This study assesses the burden of undernutrition among infants in India through a single measure “Composite index of anthropometric failure” helps to detect the multiple undernutrition along with the individual conventional three types of undernutrition like underweight, stunted, and wasted and its association with different socio-economic variables. Methods: It is a cross-sectional study. The data has been taken from the fifth national family health (NFHS-5) survey of 2019–21 and the sample size is 32, 875 of 0–11-month-old infants. Undernutrition has been assessed by the status of underweight, stunting and wasting and by CIAF methods from WHO-Z score system. The associated socio-demographic factors are sex of infants, place of residence, mother\u27s education, age and BMI, religion, wealth index of the family and different zones of India. To measure the intensity of association between different socio-demographic variables and undernutrition, multivariate analysis has been done. Results: The study reveals that among infants in India, the prevalence of underweight, stunting, and wasting are 23.4%, 24.3% 23.7%, respectively, and prevalence of total undernutrition through CIAF method is 47.4%. The result shows that the percentage of underweight and stunting are increasing with age whereas the percentage of wasting and CIAF are decreasing with age. The data also shows that CIAF is inversely related with urban place of residence, baby girl, mother\u27s education, age and BMI levels, “Christian and other” religions, wealth index of the family and birthweight of babies
On Cumulative Information Measures: Properties, Inference and Applications
No full textIn this thesis, various weighted information measures based on cumulative distribution functions and survival functions of the underlying random variables are proposed and their properties are studied. Dynamic information measures are introduced which are defined in terms of the residual and past lifetimes of the underlying random variables. Aging classes based on the dynamic information measures are discussed and characterization results for Rayleigh and power distributions are obtained. Non-parametric estimators of these measures are proposed using empirical distribution function, L-Statistics and Kernel function. Asymptotic properties of these estimators are investigated. Exponentiality tests for complete and censored data and uniformity tests are developed as applications. Also Applications of cumulative residual extropy measure in reliability engineering and hypothesis testing problems are discussed. Optimal designs for progressive Type-II censored experiments using cumulative entropy measures are investigated. Numerous examples are provided throughout the course of this thesis for illustrations
Limiting spectral distribution of some patterned random matrices with independent entries
No full textThe scaled standard Wigner matrix and its limiting eigenvalue distribution, namely the semi-circular distribution, has attracted much attention. The 2kth moment of the limit equals the number of non-crossing pair-partitions of the set f1; 2; : : : ; 2kg. There are several extensions of this result in the literature. Patterned random matrices such as the reverse circulant, the symmetric circulant, the Toeplitz and the Hankel matrices and their almost sure limiting spectral distribution (LSD), have also been studied quite extensively. Under the assumption that the entries are taken from an i.i.d. sequence with finite variance, the LSD are tied together by a common thread—the 2kth moment of the limit equals a weighted sum over di erent types of pair-partitions of the set f1; 2; : : : ; 2kg and are universal. Some results are also known for the sparse case. We discuss extension of these results by relaxing significantly the i.i.d. assumption. With suitable assumption on the entries, the limits are defined via a larger class of partitions and are also not universal. In the process we show how some new sets of partitions gain importance in finding the moments of the limits. Several existing and new results forWigner and other patterned matrices, their band and sparse versions, as well as for matrices with continuous and discrete variance profile follow as special cases. We also look into similar generalisation of existing results for XXT matrices with independent entries that satisfy certain moment conditions. In this case too, some new results are generated besides obtaining the existing results as special cases