arXiv subjects

Shabnam Sodagari

Publications and source records attributed to Shabnam Sodagari.

9 recordsLinked to original sources

arXiv2026

Optimal FALQON for Quantum Approximate Optimization via Layer-wise Parameter Tuning

Feedback-based adaptive quantum optimization (FALQON) is a promising approach for solving combinatorial problems on noisy intermediate-scale quantum (NISQ) devices, requiring only single circuit evaluations per layer. However, standard FALQON relies on fixed hyperparameters that severely limit convergence speed, requiring hundreds to thousands of layers for acceptable solutions. This paper proposes Optimal FALQON, an optimization-based formulation that treats the per-layer time step ($δ_k$) and scaling factor ($M_k$) as decision variables optimized via classical methods. We present a comprehensive empirical study on all 94 non-isomorphic 3-regular graphs with 12 vertices, comparing Optimal FALQON with standard FALQON and multiple QAOA variants. Results demonstrate statistically significant improvements in success probability, evaluation efficiency, and depth-normalized cost across the evaluated benchmarks. Furthermore, initializing QAOA with parameters from Optimal FALQON yields superior warm-start performance compared to fixed initialization.

quant-ph↗

arXiv2026

Recovering Candidate Circadian Regulators of Arrhythmic Pituitary Hormone Genes Using Reliability-Weighted Magnetic Laplacian with rwMagLap

We study how to recover candidate circadian-clock regulators of pituitary hormone genes that are important for women's health but do not show a clear 24-hour rhythm in bulk tissue, aiming to nominate clock-linked regulatory targets that could inform future chronopharmacologic and chronotherapeutic strategies. We propose \textbf{rwMagLap}, which builds a graph on rhythmic backbone genes. For each edge, we combine 24-hour fit quality with peak-time phase, represented as a complex unit-circle value, yielding a Hermitian adjacency matrix and a magnetic Laplacian. We insert arrhythmic hormone genes, treated as anchors, by a reliability-weighted nearest-neighbor projection. The projected anchor-neighbor weights are pooled into a soft teleport distribution, and complex personalized PageRank then ranks rhythmic backbone genes by the magnitude of their PageRank scores. In pituitary data, we find that all 11 women's-health anchors are arrhythmic. Even so, we find that the top-50 list is $7.95\times$ enriched for the 13-gene KEGG circadian set (7 of the 8 set genes in the 454-gene backbone; corrected Benjamini-Hochberg (BH) $p_{\mathrm{BH}}=4\times10^{-6}$) and $4.54\times$ enriched for the 111-gene Reactome set (8 of 16 genes; $p_{\mathrm{BH}}=1.6\times10^{-4}$), while a phase-blind real-valued baseline recovers none. We recover candidates through reliability weighting and phase-aware seeding rather than through magnetic propagation. The magnetic phase adds a different capability: it represents temporal order. On pituitary backbone, the magnetic embedding recovers measured peak-time order of connected pituitary genes with accuracy $0.971$, while $q{=}0$, i.e., no magnetic charge, is at chance.

q-bio.QM↗

arXiv2026

Quantum Causal Discovery via Amplitude Estimation of Kullback-Leibler Divergence

Causal discovery from observational data underpins applications in finance, climate modeling, and machine learning. Constraint-based causal discovery reduces structure learning to a sequence of conditional independence (CI) tests, where each test decides independence by estimating conditional mutual information $I(X;Y \mid Z)$ to additive precision $τ$ and thresholding against it. Classically this requires $Θ(1/τ^{2})$ samples per test, a cost that dominates in the high-precision regime typical of weak dependencies. We present QKLA (Quantum Kullback--Leibler Amplitude estimation), a quantum algorithm that encodes a clipped log-density ratio as a bounded amplitude and applies amplitude estimation to recover a clipped KL expectation. Given coherent oracle access to the relevant distributions and a reversible log-ratio arithmetic oracle, QKLA achieves a quadratic precision improvement, needing only $\mathcal{O}((L/τ)\log(1/δ))$ queries, where $L$ is the log-ratio clip bound. Under per-stratum conditional-oracle access and a margin assumption for CI decisions, embedding this estimator in the PC algorithm compounds to an $\widetildeΩ(1/(Lτ))$ reduction in total oracle queries. We validate the theory in three experiments. A gate-level state-vector simulation of the full QKLA circuit confirms the predicted $\mathcal{O}(1/M)$ error decay. Across $K=20$ random binary distributions, classical and quantum error scalings match theory to within $0.01$ in slope. In an oracle-model benchmark inside PC on two networks (\textsc{Asia}, 8 nodes; \textsc{Synthetic-12}, 12 nodes), the quantum CI subroutine reaches comparable skeleton-recovery $F_1$ while using $2.7$--$3.2\times$ fewer oracle queries at $τ= 5\cdot 10^{-3}$ bits and $4.0$--$7.4\times$ fewer at $τ= 10^{-3}$ bits.

quant-ph↗

arXiv2026

Leveraging Quantum-Based Architectures for Robust Diagnostics

Quantum machine learning has emerged as a promising approach for medical image analysis, particularly in settings where compact models and expressive feature representations are desired. This paper presents a hybrid classical--quantum diagnostic framework that integrates dataset-specific preprocessing, transfer learning, and quantum convolutional neural networks (QCNNs) for multi-class medical image classification. This approach is evaluated on three distinct tasks: kidney disease diagnosis from computed tomography images, cervical cell classification from pap smear images, and brain tumor classification from magnetic resonance imaging. For each dataset, a pretrained encoder is used to extract latent features, which are then embedded into quantum states through angle or amplitude encoding and processed by a QCNN. Experimental results show strong and stable convergence across all datasets. The proposed hybrid models achieve 99% test accuracy on kidney CT classification, 97% on cervical cell classification, and 99% on brain tumor classification. In comparative evaluations for precision, recall, and F1, the hybrid QCNN models consistently outperform classical CNN baselines using the same pretrained encoders and similar hyperparameter settings, while requiring fewer trainable parameters. These results demonstrate the potential of quantum-enhanced architectures for robust and efficient medical diagnostics.

cs.CV↗

arXiv2019

Deep UL2DL: Channel Knowledge Transfer from Uplink to Downlink

Knowledge of the channel state information (CSI) at the transmitter side is one of the primary sources of information that can be used for the efficient allocation of wireless resources. Obtaining downlink (DL) CSI in Frequency Division Duplexing (FDD) systems from uplink (UL) CSI is not as straightforward as in TDD systems. Therefore, users usually feed the DL-CSI back to the transmitter. To remove the need for feedback (and thus having less signaling overhead), we propose to use two recent deep neural network structures, i.e., convolutional neural networks and generative adversarial networks (GANs) to infer the DL-CSI by observing the UL-CSI. The core idea of our data-driven scheme is exploiting the fact that both DL and UL channels share the same propagation environment. As such, we extracted the environment information from the UL channel response to a latent domain and then transferred the derived environment information from the latent domain to predict the DL channel. To overcome incorrect latent domain and the problem of oversimplistic assumptions, in this work, we did not use any specific parametric model and instead used data-driven approaches to discover the underlying structure of data without any prior model assumptions. To overcome the challenge of capturing the UL-DL joint distribution, we used a mean square error-based variant of the GAN structure with improved convergence properties called boundary equilibrium GAN (BEGAN). For training and testing we used simulated data of Extended Vehicular-A (EVA) and Extended Typical Urban (ETU) models. Simulation results verified that our methods can accurately infer and predict the downlink CSI from the uplink CSI for different multipath environments in FDD communications.

eess.SP↗

arXiv2019

Underpinnings of User-Channel Allocation in Non-Orthogonal Multiple Access for 5G

Non-orthogonal multiple access (NOMA) is a part of 5th generation (5G) communication systems. This article presents the underpinnings and underlying structures of the problem of NOMA user-channel allocation. Unlike the heuristics for NOMA user-channel allocation, the presented results are guaranteed to converge to a solution. In addition, the solutions are stable. Generally, the results apply to any NOMA system. Unlike the orthogonal frequency division multiple access (OFDMA) resource allocation problem, the core matching is not the solution to NOMA resource allocation. The conditions under which the fix-point NOMA resource allocation is guaranteed to be stable from the viewpoint of both the base station and the NOMA users are described. In addition, relationships of NOMA user-channel resource allocation to game models and subgame perfect Nash equilibria are elucidated.

eess.SP↗

arXiv2014

On Effects of Imperfect Channel State Information on Null Space Based Cognitive MIMO Communication

In cognitive radio networks, when secondary users transmit in the null space of their interference channel with primary user, they can avoid interference. However, performance of this scheme depends on knowledge of channel state information for secondary user to perform inverse waterfilling. We evaluate the effects of imperfect channel estimation on error rates and performance degradation of primary user and elucidate the tradeoffs, such as amount of interference and guard distance. Results show that, based on the amount of perturbation in channel matrices, performance of null space based technique can degrade to that of open loop MIMO. Outcomes presented in this paper also apply to null space based MIMO radar waveform design to avoid interference with commercial communication systems, operating in same or adjacent bands.

cs.IT↗

arXiv2014

Application of Asynchronous Weak Commitment Search in Autonomous Quality of Service Provision in Cognitive Radio Networks

This article presents a distributed solution to autonomous quality of service provision in cognitive radio networks. Specifically, cognitive STDMA and CDMA communication networks are studied. Based on asynchronous weak commitment search the task of QoS provision is distributed among different network nodes. Simulation results verify this scheme converges very fast to optimal solution, which makes it suitable for practical real time systems. This application of artificial intelligence in wireless and mobile communications can be used in home automation and networking, and vehicular technology. The generalizations and extensions of this approach can be used in Long Term Evolution Self Organizing Networks (LTE-SONs). In addition, it can pave the way for decentralized and autonomous QoS provision in capillary networks that reach end nodes at Internet of Things, where central management is either unavailable or not efficient.

cs.NI↗

arXiv2010

Error Concealment in Image Communication Using Edge Map Watermarking and Spatial Smoothing

We propose a novel error concealment algorithm to be used at the receiver side of a lossy image transmission system. Our algorithm involves hiding the edge map of the original image at the transmitter within itself using a robust watermarking scheme. At the receiver, wherever a lost block is detected, the extracted edge information is used as border constraint for the spatial smoothing employing the intact neighboring blocks in order to conceal errors. Simulation results show the superiority of our technique over existing methods even in case of high packet loss ratios in the communication network.

cs.MM↗