arXiv subjects

Adnan Khan

Publications and source records attributed to Adnan Khan.

12 recordsLinked to original sources

arXiv2026

TactileEval: A Step Towards Automated Fine-Grained Evaluation and Editing of Tactile Graphics

Tactile graphics require careful expert validation before reaching blind and visually impaired (BVI) learners, yet existing datasets provide only coarse holistic quality ratings that offer no actionable repair signal. We present TactileEval, a three-stage pipeline that takes a first step toward automating this process. Drawing on expert free-text comments from the TactileNet dataset, we establish a five-category quality taxonomy; encompassing view angle, part completeness, background clutter, texture separation, and line quality aligned with BANA standards. We subsequently gathered 14,095 structured annotations via Amazon Mechanical Turk, spanning 66 object classes organized into six distinct families. A reproducible ViT-L/14 feature probe trained on this data achieves 85.70% overall test accuracy across 30 different tasks, with consistent difficulty ordering suggesting the taxonomy suggesting the taxonomy captures meaningful perceptual structure. Building on these evaluations, we present a ViT-guided automated editing pipeline that routes classifier scores through family-specific prompt templates to produce targeted corrections via gpt-image-1 image editing. Code, data, and models are available at https://TactileEval.github.io/

cs.CV↗

arXiv2026

Potential Role of Agentic Artificial Intelligence in Toxicologic Pathology

As the volume and complexity of nonclinical toxicology studies continue to increase, toxicologic pathology reporting faces persistent challenges, including fragmented sources of data (e.g., histopathology images, clinical pathology and other study data, adverse effects database, mechanistic literature), variable reporting timelines and heightened regulatory expectations. This white paper examines the emerging role of agentic artificial intelligence (AI) in addressing these issues through coordinated workflow orchestration, data integration, and pathologist-in-the-loop report generation. Based on a closed-door roundtable held during the 2025 Society of Toxicologic Pathology (STP) Annual Meeting and follow-on discussions, this paper synthesizes the perspectives of leading toxicologic pathologists, toxicologists, and AI developers. It outlines the key pain points in current reporting workflows, identifies realistic near-term use cases for agentic AI, and describes major adoption barriers including requirements for transparency, validation, and organizational readiness. A phased adoption roadmap and pilot design considerations are proposed to help support responsible evaluation and deployment of agentic AI system in nonclinical settings. The paper concludes by emphasizing the need for coordinated efforts across pharmaceutical organizations, CROs, academia, and regulators to establish shared standards, benchmarks, and governance frameworks that will lead to safe, transparent, and trustworthy integration of AI into toxicologic science.

cs.CY↗

arXiv2025

Mathematical Modeling of Biofilm Eradication Using Optimal Control

We propose and analyze a model for antibiotic resistance transfer in a bacterial biofilm and examine antibiotic dosing strategies that are effective in bacterial elimination. In particular, we consider a 1-D model of a biofilm with susceptible, persistor and resistant bacteria. Resistance can be transferred to the susceptible bacteria via horizontal gene transfer (HGT), specifically via conjugation. We analyze some basic properties of the model, determine the conditions for existence of disinfection and coexistence states, including boundary equilibria and their stability. Numerical simulations are performed to explore different modeling scenarios and support our theoretical findings. Different antibiotic dosing strategies are then studied, starting with a continuous dosing; here we note that high doses of antibiotic are needed for bacterial elimination. We then consider periodic dosing, and again observe that insufficient levels of antibiotic per dose may lead to treatment failure. Finally, using an extended version of Pontryagin's maximum principle we determine efficient antibiotic dosing protocols, which ensure bacterial elimination while keeping the total dosing low; we note that this involves a tapered dosing which reinforces results presented in other clinical and modeling studies. We study the optimal dosing for different parameter values and note that the optimal dosing schedule is qualitatively robust.

q-bio.PE↗

arXiv2025

TactileNet: Bridging the Accessibility Gap with AI-Generated Tactile Graphics for Individuals with Vision Impairment

Tactile graphics are essential for providing access to visual information for the 43 million people globally living with vision loss. Traditional methods for creating these graphics are labor-intensive and cannot meet growing demand. We introduce TactileNet, the first comprehensive dataset and AI-driven framework for generating embossing-ready 2D tactile templates using text-to-image Stable Diffusion (SD) models. By integrating Low-Rank Adaptation (LoRA) and DreamBooth, our method fine-tunes SD models to produce high-fidelity, guideline-compliant graphics while reducing computational costs. Quantitative evaluations with tactile experts show 92.86% adherence to accessibility standards. Structural fidelity analysis revealed near-human design similarity, with an SSIM of 0.538 between generated graphics and expert-designed tactile images. Notably, our method preserves object silhouettes better than human designs (SSIM = 0.259 vs. 0.215 for binary masks), addressing a key limitation of manual tactile abstraction. The framework scales to 32,000 images (7,050 high-quality) across 66 classes, with prompt editing enabling customizable outputs (e.g., adding or removing details). By automating the 2D template generation step-compatible with standard embossing workflows-TactileNet accelerates production while preserving design flexibility. This work demonstrates how AI can augment (not replace) human expertise to bridge the accessibility gap in education and beyond. Code, data, and models will be publicly released to foster further research.

cs.CV↗

arXiv2025

Will artificial intelligence accelerate or delay the race between nuclear energy technology budgeting and net-zero emissions?

This study explores the impact of nuclear energy technology budgeting and artificial intelligence on carbon dioxide (CO2) emissions in 20 OECD economies. Unlike previous research that relied on conventional panel techniques, we utilize the Method of Moment Quantile Regression panel data estimation techniques. This approach provides quantile-specific insights while addressing issues of endogeneity and heteroscedasticity, resulting in a more nuanced and robust understanding of complex relationships. A novel aspect of this research work is introducing the moderating effect of artificial intelligence on the relationship between nuclear energy and CO2 emissions. The results found that the direct impact of artificial intelligence on CO2 emissions is significant, while the effect of nuclear energy technology budgeting is not. Additionally, artificial intelligence moderates the relationship between nuclear energy technology budgeting and CO2 emissions, aiding nuclear energy in reducing carbon emissions across OECD countries. Our findings indicate that transitioning to a low-carbon future is achievable by replacing fossil fuel energy sources with increased integration of artificial intelligence to promote nuclear energy technologies. This study demonstrates that energy innovations can serve as effective climate-resilience strategies to mitigate the impacts of climate change.

econ.GN↗

arXiv2025

Articulating the role of nuclear energy in the circular economy of China: A machine learning approach

Nuclear energy is increasingly recognized as a critical component of circular economy frameworks due to its capacity to provide a stable, low-carbon energy source. Reducing dependency on fossil fuels promotes sustainable practices and aligns with circular economy goals such as resource efficiency, pollution reduction, and waste minimization. The existing literature has primarily focused on the contribution of nuclear energy to decarbonization, whereas the potential of nuclear energy in facilitating a circular economy has been largely neglected. In light of this context, this paper explores the impact of nuclear energy on the circular economy, thereby offering strong econometric evidence. The study used the advanced econometric tool Dynamic Auto-Regressive Distributive Lag (DYNARDL) method for empirical estimation to obtain long- and short-run estimates. The regression estimates, derived from a sample of China spanning 1990 to 2017, support the hypothesis that nuclear energy negatively impacts the circular economy in both the long- and short-run. Advanced econometric tests confirm the stability of the models, homoscedasticity, and the absence of serial correlation, ensuring the reliability of our findings. The study emphasizes the importance of policy strategies, including expanding nuclear energy adoption, advancing environmental technologies, and the effective use of nuclear energy by integrating comprehensive datasets and methodologies; this paper provides a foundation for scalable and equitable solutions as China moves toward a greener and more sustainable future.

econ.GN↗

arXiv2024

A Fault Prognostic System for the Turbine Guide Bearings of a Hydropower Plant Using Long-Short Term Memory (LSTM)

Hydroelectricity, being a renewable source of energy, globally fulfills the electricity demand. Hence, Hydropower Plants (HPPs) have always been in the limelight of research. The fast-paced technological advancement is enabling us to develop state-of-the-art power generation machines. This has not only resulted in improved turbine efficiency but has also increased the complexity of these systems. In lieu thereof, efficient Operation & Maintenance (O&M) of such intricate power generation systems has become a more challenging task. Therefore, there has been a shift from conventional reactive approaches to more intelligent predictive approaches in maintaining the HPPs. The research is therefore targeted to develop an artificially intelligent fault prognostics system for the turbine bearings of an HPP. The proposed method utilizes the Long Short-Term Memory (LSTM) algorithm in developing the model. Initially, the model is trained and tested with bearing vibration data from a test rig. Subsequently, it is further trained and tested with realistic bearing vibration data obtained from an HPP operating in Pakistan via the Supervisory Control and Data Acquisition (SCADA) system. The model demonstrates highly effective predictions of bearing vibration values, achieving a remarkably low RMSE.

cs.AI↗

arXiv2024

MedPromptX: Grounded Multimodal Prompting for Chest X-ray Diagnosis

Chest X-ray images are commonly used for predicting acute and chronic cardiopulmonary conditions, but efforts to integrate them with structured clinical data face challenges due to incomplete electronic health records (EHR). This paper introduces MedPromptX, the first clinical decision support system that integrates multimodal large language models (MLLMs), few-shot prompting (FP) and visual grounding (VG) to combine imagery with EHR data for chest X-ray diagnosis. A pre-trained MLLM is utilized to complement the missing EHR information, providing a comprehensive understanding of patients' medical history. Additionally, FP reduces the necessity for extensive training of MLLMs while effectively tackling the issue of hallucination. Nevertheless, the process of determining the optimal number of few-shot examples and selecting high-quality candidates can be burdensome, yet it profoundly influences model performance. Hence, we propose a new technique that dynamically refines few-shot data for real-time adjustment to new patient scenarios. Moreover, VG narrows the search area in X-ray images, thereby enhancing the identification of abnormalities. We also release MedPromptX-VQA, a new in-context visual question answering dataset encompassing interleaved images and EHR data derived from MIMIC-IV and MIMIC-CXR-JPG databases. Results demonstrate the SOTA performance of MedPromptX, achieving an 11% improvement in F1-score compared to the baselines. Code and data are publicly available on https://github.com/BioMedIA-MBZUAI/MedPromptX.

cs.CV↗

arXiv2024

Fine-Tuned Large Language Models for Symptom Recognition from Spanish Clinical Text

The accurate recognition of symptoms in clinical reports is significantly important in the fields of healthcare and biomedical natural language processing. These entities serve as essential building blocks for clinical information extraction, enabling retrieval of critical medical insights from vast amounts of textual data. Furthermore, the ability to identify and categorize these entities is fundamental for developing advanced clinical decision support systems, aiding healthcare professionals in diagnosis and treatment planning. In this study, we participated in SympTEMIST, a shared task on the detection of symptoms, signs and findings in Spanish medical documents. We combine a set of large language models fine-tuned with the data released by the organizers.

cs.CL↗

arXiv2024

Improving Pseudo-labelling and Enhancing Robustness for Semi-Supervised Domain Generalization

Beyond attaining domain generalization (DG), visual recognition models should also be data-efficient during learning by leveraging limited labels. We study the problem of Semi-Supervised Domain Generalization (SSDG) which is crucial for real-world applications like automated healthcare. SSDG requires learning a cross-domain generalizable model when the given training data is only partially labelled. Empirical investigations reveal that the DG methods tend to underperform in SSDG settings, likely because they are unable to exploit the unlabelled data. Semi-supervised learning (SSL) shows improved but still inferior results compared to fully-supervised learning. A key challenge, faced by the best-performing SSL-based SSDG methods, is selecting accurate pseudo-labels under multiple domain shifts and reducing overfitting to source domains under limited labels. In this work, we propose new SSDG approach, which utilizes a novel uncertainty-guided pseudo-labelling with model averaging (UPLM). Our uncertainty-guided pseudo-labelling (UPL) uses model uncertainty to improve pseudo-labelling selection, addressing poor model calibration under multi-source unlabelled data. The UPL technique, enhanced by our novel model averaging (MA) strategy, mitigates overfitting to source domains with limited labels. Extensive experiments on key representative DG datasets suggest that our method demonstrates effectiveness against existing methods. Our code and chosen labelled data seeds are available on GitHub: https://github.com/Adnan-Khan7/UPLM

cs.CV↗

arXiv2022

Baryonic Post-Processing of N-body Simulations, with Application to Fast Radio Bursts

Where the cosmic baryons lie in and around galactic dark matter halos is only weakly constrained. We develop a method to quickly paint on models for their distribution. Our approach uses the statistical advantages of $N$-body simulations, while painting on the profile of gas around individual halos in ways that can be motivated by semi-analytic models or zoom-in hydrodynamic simulations of galaxies. Possible applications of the algorithm include extragalactic dispersion measures to fast radio bursts (FRBs), the Sunyaev-Zeldovich effect, baryonic effects on weak lensing, and cosmic metal enrichment. As an initial application, we use this tool to investigate how the baryonic profile of foreground galactic-mass halos affects the statistics of the dispersion measure (DM) towards cosmological FRBs. We show that the distribution of DM is sensitive to the distribution of baryons in galactic halos, with viable gas profile models having significantly different probability distributions for DM to a given redshift. We also investigate the requirements to statistically measure the circumgalactic electron profile for FRB analyses that stack DM with impact parameter to foreground galaxies, quantifying the size of the contaminating "two-halo" term from correlated systems and the number of FRBs for a high significance detection. Publicly available Python modules implement our CGMBrush algorithm.

astro-ph.CO↗

arXiv2017

Spin-optomechanical coupling between light and a nanofiber torsional mode

Light that carries linear or angular momentum can interact with a mechanical object giving rise to optomechanical effects. In particular, a photon transfers its intrinsic angular momentum to an object when the object either absorbs the photon or changes the photon polarization, as in an action/reaction force pair. Here, we present the implementation of light-induced selective resonant driving of the torsional mechanical modes of a single-mode tapered optical nanofiber. The nanofiber torsional mode spectrum is characterized by polarimetry, showing narrow natural resonances (Q$\approx$2,000). By sending amplitude modulated light through the nanofiber, we resonantly drive individual torsional modes as a function of the light polarization. By varying the input polarization to the fiber, we find the largest amplification of a mechanical oscillation (>35 dB) is observed when driving the system with light containing longitudinal spin on the nanofiber waist. These results present optical nanofibers as a platform suitable for quantum spin-optomechanics experiments.

physics.optics↗