arXiv subjects

Jaskaran Singh

Publications and source records attributed to Jaskaran Singh.

At least 19 recordsLinked to original sources

arXiv2026

Optimal Stratified Allocation for Rare-Event Onset Forecasting in Dependent Sequences

Let a finite population of n labelled examples carry a class-weighted loss, with pi*n in a rare positive class weighted by N0/N1. We study estimation of total risk from a subsample K << n under designs allocating K0 and K1 draws to the two strata. We derive the exact finite-population variance of the weighted risk estimator under class-conditional sampling without replacement and solve for the optimal allocation. The class multiplier inflates positive-stratum dispersion by the imbalance ratio, causing that ratio to cancel from the optimal allocation and making equal, rather than proportional, allocation the natural default. Simple random sampling is dominated by an explicit between-stratum term; an exact bias identity shows that cluster-representative selection has no general unbiasedness guarantee; and a Serfling bound transfers the allocation result to selection error over a finite candidate set. Under the implemented truncation, the realised allocation ratio is gamma=min(2*pi/f,1), independent of n, yielding the parameter-free efficiency prediction A(pi,f)=gamma/[pi(1-pi)(1+gamma)^2]. A separate measurability result bounds the record occupied by a labelled example, determining the required train-test separation and controlling departure from block independence under absolute regularity. We test these predictions on forecasting the onset of statistically explosive price regimes, dated ex post by the Phillips-Shi-Yu procedure, using 350 U.S. equities from 2004-2011 with under 1% positive rows and five purged forward blocks. The predicted ordering of the four designs holds, and at the 10-day horizon the five design points are ordered exactly as predicted by A (Spearman rho=1, exact p=0.0167). The predicted dependence on pi across horizons does not hold; we identify the channels lying outside the design-based argument.

math.ST↗

arXiv2026

Efficient circuit compression by multiqudit entangling gates in linear optical quantum computation

Linear optical quantum computation (LOQC) offers a promising platform for scalable quantum information processing, but its scalability is fundamentally constrained by the probabilistic nature of non-local entangling gates. Qudit circuit compression schemes mitigate this issue by encoding multiple qubits onto qudits. However, these schemes become inefficient when only a subset of the encoded qubits is required to participate in the non-local entangling gate, leading to an exponential increase in the number of non-local gates. In this Letter, we address this bottleneck by demonstrating the existence of multi-level control-Z (CZ) gates for qudits encoded in multiple spatial modes in LOQC. Unlike conventional two-level CZ gates, which act only on a single pair of modes, multi-level CZ gates impart a conditional phase shift for an arbitrarily chosen subset of the spatial modes. We present two explicit linear optical schemes that realize such operations, illustrating a fundamental trade-off between prior information about the input quantum state and the physical resources required. The first scheme is realized with a constant success probability of $1/8$ independent of the qudit dimension using a single non-local entangling gate, at the cost of state dependence, which is significantly better than the current success probability of $1/9$. Our second scheme provides a fully state independent realization reducing the number of non-local gates to $\mathcal{O}(2^{r_1}+2^{r_2})$ as compared to the existing bound of $\mathcal{O}(2^{r_1+r_2})$ where $r_1$ and $r_2$ are the number of qubits to be removed as control in the qudits. The success probability of the realization is $\frac{1}{2} \left(\frac{1}{8}\right)^{2^{r_1}+2^{r_2}}$. When combined with qudit circuit compression schemes, our results improve upon a key scalability limitation and significantly improve the efficiency of LOQC architectures.

quant-ph↗

arXiv2025

MahaTTS: A Unified Framework for Multilingual Text-to-Speech Synthesis

Current Text-to-Speech models pose a multilingual challenge, where most of the models traditionally focus on English and European languages, thereby hurting the potential to provide access to information to many more people. To address this gap, we introduce MahaTTS-v2 a Multilingual Multi-speaker Text-To-Speech (TTS) system that has excellent multilingual expressive capabilities in Indic languages. The model has been trained on around 20K hours of data specifically focused on Indian languages. Our approach leverages Wav2Vec2.0 tokens for semantic extraction, and a Language Model (LM) for text-to-semantic modeling. Additionally, we have used a Conditional Flow Model (CFM) for semantics to melspectogram generation. The experimental results indicate the effectiveness of the proposed approach over other frameworks. Our code is available at https://github.com/dubverse-ai/MahaTTSv2

eess.AS↗

arXiv2025

Tuning electronic correlations in the Kagome metals $RT_3$B$_2$

The $RT_3$B$_2$ ($R=$Y, Lu, $T=$ Co, Os) family hosts a perfect kagome lattice of $T$ atoms, offering an interesting platform to investigate the interplay of electronic structure, superconductivity, and lattice dynamics. Here, we compare two members of this family, LuOs$_3$B$_2$ and YCo$_3$B$_2$, with similar crystallography but differing chemical composition, leading to distinct electronic correlation strengths and spin-orbit coupling effects. We confirm superconductivity in LuOs$_3$B$_2$ with $T_c = 4.75$K, while YCo$_3$B$_2$ remains non-superconducting above 1.8K. First-principles estimates of the electron-phonon coupling for LuOs$_3$B$_2$ are consistent with its observed $T_c$ and suggest a moderate coupling strength. Both materials exhibit kagome-derived electronic features, including quasi-flat bands, Dirac cones, and van Hove singularities. Fermi surface calculations reveal quasi-one-dimensional behavior along the $c$-axis in YCo$_3$B$_2$, in contrast to the more three-dimensional Fermiology of LuOs$_3$B$_2$. Phonon calculations for LuOs$_3$B$_2$ show imaginary modes, indicating potential lattice instabilities. Experimental estimates of the Wilson and Kadowaki-Woods ratios point to non-negligible electronic correlations in both compounds.

cond-mat.str-el↗

arXiv2025

Train Once, Forget Precisely: Anchored Optimization for Efficient Post-Hoc Unlearning

As machine learning systems increasingly rely on data subject to privacy regulation, selectively unlearning specific information from trained models has become essential. In image classification, this involves removing the influence of particular training samples, semantic classes, or visual styles without full retraining. We introduce \textbf{Forget-Aligned Model Reconstruction (FAMR)}, a theoretically grounded and computationally efficient framework for post-hoc unlearning in deep image classifiers. FAMR frames forgetting as a constrained optimization problem that minimizes a uniform-prediction loss on the forget set while anchoring model parameters to their original values via an $\ell_2$ penalty. A theoretical analysis links FAMR's solution to influence-function-based retraining approximations, with bounds on parameter and output deviation. Empirical results on class forgetting tasks using CIFAR-10 and ImageNet-100 demonstrate FAMR's effectiveness, with strong performance retention and minimal computational overhead. The framework generalizes naturally to concept and style erasure, offering a scalable and certifiable route to efficient post-hoc forgetting in vision models.

cs.LG↗

arXiv2024

Local contextuality-based self-tests are sufficient for randomness expansion secure against quantum adversaries

In quantum cryptography, secure randomness expansion involves using a short private string of random bits to generate a longer one, even in the presence of an adversary who may have access to quantum resources. In this work, we demonstrate that local contextuality-based self-tests are sufficient to construct a randomness expansion protocol that is secure against computationally unbounded quantum adversaries. Our protocol is based on self-testing from non-contextuality inequalities and we prove that our scheme asymptotically produces secure random numbers which are $\mathcal{O}(m\sqrt{\epsilon})$-close to uniformly distributed and private, where $\epsilon$ is the robustness parameter of the self-test and $m$ is the length of the generated random bit string. Our protocol is semi-device-independent in the sense that it inherits any assumptions necessary for the underlying self-test.

quant-ph↗

arXiv2024

Analyzing LLM Usage in an Advanced Computing Class in India

This study examines the use of large language models (LLMs) by undergraduate and graduate students for programming assignments in advanced computing classes. Unlike existing research, which primarily focuses on introductory classes and lacks in-depth analysis of actual student-LLM interactions, our work fills this gap. We conducted a comprehensive analysis involving 411 students from a Distributed Systems class at an Indian university, where they completed three programming assignments and shared their experiences through Google Form surveys. Our findings reveal that students leveraged LLMs for a variety of tasks, including code generation, debugging, conceptual inquiries, and test case creation. They employed a spectrum of prompting strategies, ranging from basic contextual prompts to advanced techniques like chain-of-thought prompting and iterative refinement. While students generally viewed LLMs as beneficial for enhancing productivity and learning, we noted a concerning trend of over-reliance, with many students submitting entire assignment descriptions to obtain complete solutions. Given the increasing use of LLMs in the software industry, our study highlights the need to update undergraduate curricula to include training on effective prompting strategies and to raise awareness about the benefits and potential drawbacks of LLM usage in academic settings.

cs.HC↗

arXiv2023

Deterministic generation of hybrid entangled states using quantum walks

In recent times, hybrid-entanglement (HE) between a qubit and a coherent state has demonstrated superior performance in various quantum information processing tasks, particularly in quantum key distribution. Despite its theoretical advantages, efficient generation of such states in the laboratory has been a challenge. Here, we introduce a deterministic and efficient approach for generating HE states using quantum walks. Our method achieves a remarkable fidelity of $99.9\%$ with just $20$ time steps in a one-dimensional split-step quantum walk. This represents a significant improvement over prior approaches for probabilistic generation of HE states with fidelity as low as $80\%$. Our scheme not only provides a robust solution to the generation of HE states but also highlights a unique advantage of quantum walks, thereby contributing to the advancement of this burgeoning field. Moreover, our scheme is experimentally feasible with the current technology.

quant-ph↗

arXiv2023

Loophole-free Bell tests with randomly chosen subsets of measurement settings

There are bipartite quantum nonlocal correlations requiring very low detection efficiency to reach the loophole-free regime but that need too many measurement settings to be practical for actual experiments. This leads to the general problem of what can be concluded about loophole-free Bell nonlocality if only a random subset of these settings is tested. Here we develop a method to address this problem. We show that, in some cases, it is possible to detect loophole-free Bell nonlocality testing only a small random fraction of the settings. The prize to pay is a higher detection efficiency. The method allows for a novel approach to the design of loophole-free Bell tests in which, given the dimension of the local system, the visibility, and the detection efficiency available, one can calculate the fraction of the contexts needed to reach the detection-loophole-free regime. The results also enforce a different way of thinking about the costs of classically simulating quantum nonlocality, as it shows that the amount of resources that are needed can be made arbitrarily large simply by considering more contexts.

quant-ph↗

arXiv2023

VBSF-TLD: Validation-Based Approach for Soft Computing-Inspired Transfer Learning in Drone Detection

With the increasing utilization of Internet of Things (IoT) enabled drones in diverse applications like photography, delivery, and surveillance, concerns regarding privacy and security have become more prominent. Drones have the ability to capture sensitive information, compromise privacy, and pose security risks. As a result, the demand for advanced technology to automate drone detection has become crucial. This paper presents a project on a transfer-based drone detection scheme, which forms an integral part of a computer vision-based module and leverages transfer learning to enhance performance. By harnessing the knowledge of pre-trained models from a related domain, transfer learning enables improved results even with limited training data. To evaluate the scheme's performance, we conducted tests on benchmark datasets, including the Drone-vs-Bird Dataset and the UAVDT dataset. Notably, the scheme's effectiveness is highlighted by its IOU-based validation results, demonstrating the potential of deep learning-based technology in automating drone detection in critical areas such as airports, military bases, and other high-security zones.

cs.CV↗

arXiv2023

Long-distance entanglement sharing using hybrid states of discrete and continuous variables

We introduce a feasible scheme to produce high-rate long-distance entanglement which uses hybrid entanglement (HE) between continuous variables (CV) and discrete variables (DV). We show that HE can effectively remove the experimental limitations of existing CV and DV systems to produce long range entanglement. We benchmark the resulting DV entangled states using an entanglement-based quantum key distribution (EB-QKD) protocol. We show that, using HE states, EB-QKD is possible with standard telecommunication fibers for 300 km. The key idea is using the CV part, which can be adjusted to be robust against photon losses, for increasing the transmission distance, while using the DV part for achieving high secure key rates. Our results point out that HE states provide a clear advantage for practical long-distance and high-rate entanglement generation that may lead to further applications in quantum information processing.

quant-ph↗

arXiv2023

Revealing quantum contextuality using a single measurement device

In this work we analyse the notion of measurement non-contextuality (MNC) and identify contextual scenarios which involve sequential measurements of only a single measurement device. We show that any non-contextual ontological model fails to explain the statistics of outcomes of a single carefully constructed positive operator valued measure (POVM) executed sequentially on a quantum system. The context of measurement arises from the different configurations in which the device can be used. We develop an inequality from the non-contextual (NC) ontic model, and construct a quantum situation involving measurements from the KCBS inequality. We show that the resultant statistics arising from this device violate our NC inequality. This device can be generalised by incorporating measurements from arbitrary $n$-cycle contextuality inequalities of which $n = 5$ corresponds to the KCBS inequality. We show that the NC and quantum bounds for various scenarios can be derived more easily using only the functional relationships between the outcomes for larger values $n$. This makes it one of the simpler contextual inequalities to analyse.

quant-ph↗

arXiv2022

Role of electronic correlations in the Kagome lattice superconductor LaRh$_3$B$_2$

LaRh$_3$B$_2$ crystallizes in a layered structure where Rh atoms form a perfect Kagome lattice. The material shows superconductivity at $T_c \approx 2.6$~K\@ and no signature for density wave instabilities. We report our measurements of electronic transport, magnetization, and heat capacity in the normal and superconducting state, and derive normal and superconducting parameters. From first principles calculations of the electronic band structure, we identify all features of Kagome bands predominantly formed by the Rh $d$ orbitals: a flat band, Dirac cones, and van Hove singularities. The calculation of the phonon dispersions and electron-phonon coupling suggests a strong similarity between LaRh$_3$B$_2$ and AV$_3$Sb$_5$ (A=K,Cs,Rb). For LaRh$_3$B$_2$, it matches quantitatively with the observed $T_c$, supporting a conventional phonon mediated pairing mechanism. By comparison to the $A$V$_3$Sb$_5$ family, we conjecture a reduced importance of electron correlations in LaRh$_3$B$_2$.

cond-mat.supr-con↗

arXiv2022

No contextual advantage in non-paradoxical scenarios of two state vector formalism

The two state vector formalism (TSVF) was proposed by Aharonov, Bergmann, and Lebowitz (ABL) to provide a way for the counterfactual assignment of the probabilities of outcomes of contemplated but unperformed measurements on quantum systems. This formalism underlies various aspects of foundations of quantum theory and has been used significantly in the development of weak values and several proofs of quantum contextuality. We consider the application of TSVF, with pre- and post-selection (PPS) and the corresponding ABL rule, as a means to unearth quantum contextuality. We use the principle of exclusivity to classify the resultant pre- and post-selection scenarios as either paradoxical or non-paradoxical. In light of this, we find that several previous proofs of the emergence of contextuality in PPS scenarios are only possible if the principle of exclusivity is violated and are therefore classified as paradoxical. We argue that these do not constitute a proper test of contextuality. Furthermore, we provide a numerical analysis for the KCBS scenario as applied in the paradigm of TSVF and find that non-paradoxical scenarios do not offer any contextual advantage. Our approach can be easily generalized for other contextual scenarios as well.

quant-ph↗

arXiv2022

Graph-theoretic approach to Bell experiments with low detection efficiency

Bell inequality tests where the detection efficiency is below a certain threshold $\eta_{\rm{crit}}$ can be simulated with local hidden-variable models. Here, we introduce a method to identify Bell tests requiring low $\eta_{\rm{crit}}$ and relatively low dimension $d$ of the local quantum systems. The method has two steps. First, we show a family of bipartite Bell inequalities for which, for correlations produced by maximally entangled states, $\eta_{\rm{crit}}$ can be upper bounded by a function of some invariants of graphs, and use it to identify correlations that require small $\eta_{\rm{crit}}$. We present examples in which, for maximally entangled states, $\eta_{\rm{crit}} \le 0.516$ for $d=16$, $\eta_{\rm{crit}} \le 0.407$ for $d=28$, and $\eta_{\rm{crit}} \le 0.326$ for $d=32$. We also show evidence that the upper bound for $\eta_{\rm{crit}}$ can be lowered down to $0.415$ for $d=16$ and present a method to make the upper bound of $\eta_{\rm{crit}}$ arbitrarily small by increasing the dimension and the number of settings. All these upper bounds for $\eta_{\rm{crit}}$ are valid (as it is the case in the literature) assuming no noise. The second step is based on the observation that, using the initial state and measurement settings identified in the first step, we can construct Bell inequalities with smaller $\eta_{\rm{crit}}$ and better noise robustness. For that, we use a modified version of Gilbert's algorithm that takes advantage of the automorphisms of the graphs used in the first step. We illustrate its power by explicitly developing an example in which $\eta_{\rm{crit}}$ is $12.38\%$ lower and the required visibility is $14.62\%$ lower than the upper bounds obtained in the first step. The tools presented here may allow for developing high-dimensional loophole-free Bell tests and loophole-free Bell nonlocality over long distances.

quant-ph↗

arXiv2022

Implementation of discrete positive operator valued measures on linear optical systems using cosine-sine decomposition

Positive operator valued measurements (POVMs) play an important role in efficient quantum communication and computation. While optical systems are one of the strongest candidates for long distance quantum communication and information processing, efficient methods to implement POVMs in these systems are scarce. Here we propose an all-optical scheme to implement an arbitrary POVM using linear optical components on m-dimensional Hilbert space of internal degrees of freedom. Linear optical nature of the proposed scheme makes it efficient and robust. We show how the scheme can be applied for state tomography and for preparing arbitrary mixed states.

quant-ph↗

arXiv2022

Monogamy relations of entropic non-contextual inequalities and their experimental realization

We develop a theoretical framework based on a graph theoretic approach to analyze monogamous relationships of entropic non-contextuality (ENC) inequalities. While ENC inequalities are important in quantum information theory and are well studied, theoretical as well as experimental demonstration of their monogamous nature is still elusive. We provide conditions for ENC inequalities to exhibit a monogamous relationship and derive the same for general scenarios. We show that two entropic versions of the Bell-CHSH inequality acting on a tripartite scenario exhibit a monogamous relationship, for which we provide a theoretical proof as well as an experimental validation on an NMR quantum information processor. Our experimental technique to evaluate entropies has been designed to obtain information about entropies via measurement of only the expectation values of observables.

quant-ph↗

arXiv2021

Non-Gaussian operations in measurement device independent quantum key distribution

Non-Gaussian operations in continous variable (CV) quantum key distribution (QKD) have been limited to photon subtraction on squeezed vacuum states only. This is mainly due to the ease of calculating the covariance matrix representation of such states. In this paper we study the effects of general non-Gaussian operations corresponding to photon addition, catalysis and subtraction on squeezed coherent states on CV measurement device independent (MDI) QKD. We find that non-Gaussianity coupled with coherence can yield significantly longer transmission distances than without. Particularly we observe that zero photon catalysis on two mode squeezed coherent state (TMSC) is an optimial choice for CV MDI QKD, while single photon subtraction is also a good candidate; both of them offering nearly 70 km of transmission distances. We also derive a single generalized covariance matrix for the aforementioned states which will be useful in several other aspects of CV quantum information processing.

quant-ph↗