arXiv subjects

Rahul Patel

Publications and source records attributed to Rahul Patel.

At least 19 recordsLinked to original sources

arXiv2026

NeoJaundice-AI: Smartphone-Based Neonatal Jaundice Detection Using Dual-Input Deep Learning and Synthetic Augmentation

Neonatal jaundice (hyperbilirubinemia) is one of the most common conditions affecting newborns worldwide, with India alone recording roughly 15 million cases per year. Early detection is critical, yet standard diagnosis requires blood tests that are often impractical in rural clinics where laboratory facilities are limited. This paper presents NeoJaundice-AI, a smartphone-based screening system that uses photographs of a baby's skin and sclera (eye white) to estimate jaundice severity and predict serum bilirubin levels in under three seconds without requiring internet connectivity. The proposed system is built on a dual-branch EfficientNet-B0 architecture that independently processes skin and sclera images. Deep features are fused with handcrafted YCbCr color statistics to jointly perform four-class severity classification and continuous bilirubin regression. A key contribution is a synthetic jaundice generation method that simulates bilirubin-induced yellowing through controlled YCbCr channel modifications on normal neonatal skin images. This approach addresses data scarcity, particularly for severe jaundice cases and darker Indian skin tones (Fitzpatrick Types IV to VI). In addition, a skin-tone normalization module improves prediction consistency across diverse neonatal complexions. Experimental results demonstrate an overall classification accuracy of 91.8 percent, a clinical sensitivity of 93.5 percent, and a bilirubin mean absolute error of 1.4 mg/dL. After INT8 quantization and ONNX conversion, the model size is reduced to 8.3 MB while maintaining inference times below three seconds on standard Android devices. To the best of our knowledge, this is the first India-focused neonatal jaundice AI system that combines multimodal image fusion, skin-tone adaptation, synthetic data augmentation, and fully offline mobile deployment within a single framework.

cs.CV↗

arXiv2026

AnemiaVision: Non-Invasive Anemia Detection via Smartphone Imagery Using EfficientNet-B3 with TrivialAugmentWide, Mixup Augmentation, and Persistent Patient History Management

Anemia affects over one billion people globally and remains severely under-diagnosed in low-resource regions where laboratory blood tests are inaccessible. This paper presents AnemiaVision, an end-to-end web-based system for non-invasive anemia screening from smartphone photographs of the palpebral conjunctiva and fingernail beds. The proposed pipeline fine-tunes a pre-trained EfficientNet-B3 backbone with a redesigned three-layer classifier head incorporating BatchNorm, GELU activations, and high-rate Dropout (0.45/0.35). Training employs four orthogonal accuracy-boosting techniques: TrivialAugmentWide for policy-free image augmentation, RandomErasing for spatial regularisation, Mixup (alpha=0.2) for inter-class smoothing, and cosine-annealing scheduling with linear warmup. Early stopping is governed by peak validation accuracy rather than validation loss to prevent premature termination on high-variance epochs. The deployed Flask application integrates persistent patient-history management backed by PostgreSQL on Render, with an automated database-migration entrypoint ensuring zero data loss across redeploys. Ablation experiments demonstrate that accuracy-first early stopping contributes +1.6% and Mixup contributes +2.8% to final validation accuracy. Overall, the proposed system achieves a validation accuracy of 96.2% and AUC-ROC of 0.98, compared with 44.9% validation accuracy and AUC-ROC of 0.58 from the three-epoch CPU-only baseline. Sensitivity for the anemic class reaches 0.96, making the system suitable as a first-line screening tool for community health workers in rural settings. The system is publicly accessible and source code is openly available.

cs.CV↗

arXiv2026

The Last APK: Retiring Android SDK Development for Institutional Software Using Python-Django, HTMX, and a WebView Bridge

The assumption that mobile enterprise software requires native Android SDK development has persisted for over a decade, but for institutional deployments, this assumption is not merely outdated: it is economically wasteful and technically unnecessary. This paper presents a campus management system built during an internship at the Indian Institute of Technology Gandhinagar (IIT Gandhinagar), covering housekeeping task scheduling, inventory management, horticulture tracking, worker attendance, multi-stage leave workflows, and client-side photo capture with automatic compression. The core stack uses Python-Django as the backend framework and HTMX for hypermedia-driven, mobile-responsive partial DOM updates, containing zero lines of Android SDK application logic. The entire system runs as a self-hosted Docker Compose deployment with no dependency on any external cloud service. Through architectural analysis, HTTP payload measurement, and user experience evaluation with 42 campus staff, we demonstrate that the HTMX-Django approach reduces development time by approximately 54%, reduces average HTTP payload by 91% versus full-page reload, and achieves user satisfaction scores of 4.2/5.0.

cs.SE↗

arXiv2025

Mind the Gap: Comparing Model- vs Agentic-Level Red Teaming with Action-Graph Observability on GPT-OSS-20B

As the industry increasingly adopts agentic AI systems, understanding their unique vulnerabilities becomes critical. Prior research suggests that security flaws at the model level do not fully capture the risks present in agentic deployments, where models interact with tools and external environments. This paper investigates this gap by conducting a comparative red teaming analysis of GPT-OSS-20B, a 20-billion parameter open-source model. Using our observability framework AgentSeer to deconstruct agentic systems into granular actions and components, we apply iterative red teaming attacks with harmful objectives from HarmBench at two distinct levels: the standalone model and the model operating within an agentic loop. Our evaluation reveals fundamental differences between model level and agentic level vulnerability profiles. Critically, we discover the existence of agentic-only vulnerabilities, attack vectors that emerge exclusively within agentic execution contexts while remaining inert against standalone models. Agentic level iterative attacks successfully compromise objectives that completely failed at the model level, with tool-calling contexts showing 24\% higher vulnerability than non-tool contexts. Conversely, certain model-specific exploits work exclusively at the model level and fail when transferred to agentic contexts, demonstrating that standalone model vulnerabilities do not always generalize to deployed systems.

cs.AI↗

arXiv2025

Mind the Gap: Evaluating Model- and Agentic-Level Vulnerabilities in LLMs with Action Graphs

As large language models increasingly deployed into agentic systems, existing methods face critical gaps in observing, assessing, and mitigating deployment-specific risks. We present a comprehensive, observability-driven workflow: we introduce \textbf{AgentSeer}, observability tool which decomposes agentic executions into granular \emph{action-component} graphs; we use this decomposition to rigorously quantify the gap between model-level and agent-level jailbreaking risk via cross-model validation on GPT-OSS-20B and Gemini-2.0-flash with HarmBench under single-turn and iterative-refinement attacks; we leverage action-graph risk signals to automate iterative prompt hardening against direct and iterative jailbreak attacks. Stark differences is revealed between model-level and agentic-level vulnerability profiles. Model-level evaluation reveals baseline differences: GPT-OSS-20B (39.47\% ASR) versus Gemini-2.0-flash (50.00\% ASR), with both models showing susceptibility to social engineering. However, agentic-level assessment exposes agent-specific risks invisible to traditional evaluation. We discover "agentic-only" vulnerabilities that emerge exclusively in agentic contexts, with tool-calling showing 24-60\% higher ASR across both models. Cross-model analysis reveals universal agentic patterns, where agent transfer operations as highest-risk tools, with semantic pattern revealed rather than syntactic vulnerability mechanisms. Direct attack transfer from model-level to agentic contexts shows degraded performance of successful prompts (GPT-OSS-20B: 57\% human injection ASR; Gemini-2.0-flash: 28\%), while context-aware iterative attacks successfully compromise objectives that failed at model-level, confirming systematic vulnerabilities gaps. Action-based prompt improvement substantially reduces action-averaged agentic jailbreak success on GPT-OSS-20B (direct: 45.3\%

cs.CL↗

arXiv2025

LLM4VV: Evaluating Cutting-Edge LLMs for Generation and Evaluation of Directive-Based Parallel Programming Model Compiler Tests

The usage of Large Language Models (LLMs) for software and test development has continued to increase since LLMs were first introduced, but only recently have the expectations of LLMs become more realistic. Verifying the correctness of code generated by LLMs is key to improving their usefulness, but there have been no comprehensive and fully autonomous solutions developed yet. Hallucinations are a major concern when LLMs are applied blindly to problems without taking the time and effort to verify their outputs, and an inability to explain the logical reasoning of LLMs leads to issues with trusting their results. To address these challenges while also aiming to effectively apply LLMs, this paper proposes a dual-LLM system (i.e. a generative LLM and a discriminative LLM) and experiments with the usage of LLMs for the generation of a large volume of compiler tests. We experimented with a number of LLMs possessing varying parameter counts and presented results using ten carefully-chosen metrics that we describe in detail in our narrative. Through our findings, it is evident that LLMs possess the promising potential to generate quality compiler tests and verify them automatically.

cs.SE↗

arXiv2024

Heuristic Multiobjective Discrete Optimization using Restricted Decision Diagrams

Decision diagrams (DDs) have emerged as a state-of-the-art method for exact multiobjective integer linear programming. When the DD is too large to fit into memory or the decision-maker prefers a fast approximation to the Pareto frontier, the complete DD must be restricted to a subset of its states (or nodes). We introduce new node-selection heuristics for constructing restricted DDs that produce a high-quality approximation of the Pareto frontier. Depending on the structure of the problem, our heuristics are based on either simple rules, machine learning with feature engineering, or end-to-end deep learning. Experiments on multiobjective knapsack, set packing, and traveling salesperson problems show that our approach is highly effective, recovering over 85% of the Pareto frontier while achieving 2.5x speedups over exact DD enumeration on average, with very few non-Pareto solutions. The code is available at https://github.com/rahulptel/HMORDD.

cs.AI↗

arXiv2023

LEO: Learning Efficient Orderings for Multiobjective Binary Decision Diagrams

Approaches based on Binary decision diagrams (BDDs) have recently achieved state-of-the-art results for multiobjective integer programming problems. The variable ordering used in constructing BDDs can have a significant impact on their size and on the quality of bounds derived from relaxed or restricted BDDs for single-objective optimization problems. We first showcase a similar impact of variable ordering on the Pareto frontier (PF) enumeration time for the multiobjective knapsack problem, suggesting the need for deriving variable ordering methods that improve the scalability of the multiobjective BDD approach. To that end, we derive a novel parameter configuration space based on variable scoring functions which are linear in a small set of interpretable and easy-to-compute variable features. We show how the configuration space can be efficiently explored using black-box optimization, circumventing the curse of dimensionality (in the number of variables and objectives), and finding good orderings that reduce the PF enumeration time. However, black-box optimization approaches incur a computational overhead that outweighs the reduction in time due to good variable ordering. To alleviate this issue, we propose LEO, a supervised learning approach for finding efficient variable orderings that reduce the enumeration time. Experiments on benchmark sets from the knapsack problem with 3-7 objectives and up to 80 variables show that LEO is ~30-300% and ~10-200% faster at PF enumeration than common ordering strategies and algorithm configuration. Our code and instances are available at https://github.com/khalil-research/leo.

cs.AI↗

arXiv2022

Neur2SP: Neural Two-Stage Stochastic Programming

Stochastic Programming is a powerful modeling framework for decision-making under uncertainty. In this work, we tackle two-stage stochastic programs (2SPs), the most widely used class of stochastic programming models. Solving 2SPs exactly requires optimizing over an expected value function that is computationally intractable. Having a mixed-integer linear program (MIP) or a nonlinear program (NLP) in the second stage further aggravates the intractability, even when specialized algorithms that exploit problem structure are employed. Finding high-quality (first-stage) solutions -- without leveraging problem structure -- can be crucial in such settings. We develop Neur2SP, a new method that approximates the expected value function via a neural network to obtain a surrogate model that can be solved more efficiently than the traditional extensive formulation approach. Neur2SP makes no assumptions about the problem structure, in particular about the second-stage problem, and can be implemented using an off-the-shelf MIP solver. Our extensive computational experiments on four benchmark 2SP problem classes with different structures (containing MIP and NLP second-stage problems) demonstrate the efficiency (time) and efficacy (solution quality) of Neur2SP. In under 1.66 seconds, Neur2SP finds high-quality solutions across all problems even as the number of scenarios increases, an ideal property that is difficult to have for traditional 2SP solution techniques. Namely, the most generic baseline method typically requires minutes to hours to find solutions of comparable quality.

math.OC↗

arXiv2020

An Asymmetric Eclipse Seen Towards the Pre-Main Sequence Binary System V928 Tau

K2 observations of the weak-lined T Tauri binary V928 Tau A+B show the detection of a single, asymmetric eclipse which may be due to a previously unknown substellar companion eclipsing one component of the binary with an orbital period $>$ 66 days. Over an interval of about 9 hours, one component of the binary dims by around 60%, returning to its normal brightness about 5 hours later. From modeling of the eclipse shape we find evidence that the eclipsing companion may be surrounded by a disk or a vast ring system. The modeled disk has a radius of $0.9923\,\pm\,0.0005\,R_*$, with an inclination of $56.78\,\pm\, 0.03^\circ$, a tilt of $41.22\,\pm\,0.05^\circ$, an impact parameter of $-0.2506\,\pm\,0.0002\,R_*$ and an opacity of 1.00. The occulting disk must also move at a transverse velocity of $6.637\,\pm\,0.002\,R_*\,\mathrm{day}^{-1}$, which depending on whether it orbits V928 Tau A or B, corresponds to approximately 73.53 or 69.26 $\mathrm{km s}^{-1}$. A search in ground based archival data reveals additional dimming events, some of which suggest periodicity, but no unambiguous period associated with the eclipse observed by K2. We present a new epoch of astrometry which is used to further refine the orbit of the binary, presenting a new lower bound of 67 years, and constraints on the possible orbital periods of the eclipsing companion. The binary is also separated by 18" ($\sim$2250 au) from the lower mass CFHT-BD-Tau 7, which is likely associated with V928 Tau A+B. We also present new high dispersion optical spectroscopy that we use to characterize the unresolved stellar binary.

astro-ph.SR↗

arXiv2019

A learning-based algorithm to quickly compute good primal solutions for Stochastic Integer Programs

We propose a novel approach using supervised learning to obtain near-optimal primal solutions for two-stage stochastic integer programming (2SIP) problems with constraints in the first and second stages. The goal of the algorithm is to predict a "representative scenario" (RS) for the problem such that, deterministically solving the 2SIP with the random realization equal to the RS, gives a near-optimal solution to the original 2SIP. Predicting an RS, instead of directly predicting a solution ensures first-stage feasibility of the solution. If the problem is known to have complete recourse, second-stage feasibility is also guaranteed. For computational testing, we learn to find an RS for a two-stage stochastic facility location problem with integer variables and linear constraints in both stages and consistently provide near-optimal solutions. Our computing times are very competitive with those of general-purpose integer programming solvers to achieve a similar solution quality.

math.OC↗

arXiv2018

Correlated discrete data generation using adversarial training

Generative Adversarial Networks (GAN) have shown great promise in tasks like synthetic image generation, image inpainting, style transfer, and anomaly detection. However, generating discrete data is a challenge. This work presents an adversarial training based correlated discrete data (CDD) generation model. It also details an approach for conditional CDD generation. The results of our approach are presented over two datasets; job-seeking candidates skill set (private dataset) and MNIST (public dataset). From quantitative and qualitative analysis of these results, we show that our model performs better as it leverages inherent correlation in the data, than an existing model that overlooks correlation.

cs.LG↗

arXiv2018

275 Candidates and 149 Validated Planets Orbiting Bright Stars in K2 Campaigns 0-10

Since 2014, NASA's K2 mission has observed large portions of the ecliptic plane in search of transiting planets and has detected hundreds of planet candidates. With observations planned until at least early 2018, K2 will continue to identify more planet candidates. We present here 275 planet candidates observed during Campaigns 0-10 of the K2 mission that are orbiting stars brighter than 13 mag (in Kepler band) and for which we have obtained high-resolution spectra (R = 44,000). These candidates are analyzed using the VESPA package (Morton 2012, 2015b) in order to calculate their false-positive probabilities (FPP). We find that 149 candidates are validated with an FPP lower than 0.1%, 39 of which were previously only candidates and 56 of which were previously undetected. The processes of data reduction, candidate identification, and statistical validation are described, and the demographics of the candidates and newly validated planets are explored. We show tentative evidence of a gap in the planet radius distribution of our candidate sample. Comparing our sample to the Kepler candidate sample investigated by Fulton et al. (2017), we conclude that more planets are required to quantitatively confirm the gap with K2 candidates or validated planets. This work, in addition to increasing the population of validated K2 planets by nearly 50% and providing new targets for follow-up observations, will also serve as a framework for validating candidates from upcoming K2 campaigns and the Transiting Exoplanet Survey Satellite, expected to launch in 2018.

astro-ph.EP↗

arXiv2017

KELT-19Ab: A P~4.6 Day Hot Jupiter Transiting a Likely Am Star with a Distant Stellar Companion

We present the discovery of the giant planet KELT-19Ab, which transits the moderately bright $(\mathrm{V} \sim 9.9)$ A8V star TYC 764-1494-1 with an orbital period of 4.61 days. We confirm the planetary nature of the companion via a combination of radial velocities, which limit the mass to $< 4.1\,\mathrm{M_J}$ $(3\sigma)$, and a clear Doppler tomography signal, which indicates a retrograde projected spin-orbit misalignment of $\lambda = -179.7^{+3.7}_{-3.8}$ degrees. Global modeling indicates that the $\rm{T_{eff}} =7500 \pm 110\,\mathrm{K}$ host star has $\mathrm{M_*} = 1.62^{+0.25}_{-0.20}\,\mathrm{M_\odot}$ and $\mathrm{R_*} = 1.83 \pm 0.10\,\mathrm{R_\odot}$. The planet has a radius of $\mathrm{R_P}=1.91 \pm 0.11\,\mathrm{R_J}$ and receives a stellar insolation flux of $\sim 3.2\times 10^{9}\,\mathrm{erg\,s^{-1}\,cm^{-2}}$, leading to an inferred equilibrium temperature of $\rm{T_{EQ}} = \sim 1935\,\rm{K}$ assuming zero albedo and complete heat redistribution. With a $v\sin{I_*}=84.8\pm 2.0\,\mathrm{km\,s^{-1}}$, the host is relatively slowly rotating compared to other stars with similar effective temperatures, and it appears to be enhanced in metallic elements but deficient in calcium, suggesting that it is likely an Am star. KELT-19A would be the first detection of an Am host of a transiting planet of which we are aware. Adaptive optics observations of the system reveal the existence of a companion with late G9V/early K1V spectral type at a projected separation of $\approx 160\,\mathrm{AU}$. Radial velocity measurements indicate that this companion is bound. Most Am stars are known to have stellar companions, which are often invoked to explain the relatively slow rotation of the primary. In this case, the stellar companion is unlikely to have caused the tidal braking of the primary. However, it may have emplaced the transiting planetary companion via the Kozai-Lidov mechanism.

astro-ph.EP↗

arXiv2017

KELT-20b: A giant planet with a period of P~ 3.5 days transiting the V~ 7.6 early A star HD 185603

We report the discovery of KELT-20b, a hot Jupiter transiting a V~7.6 early A star with an orbital period of P~3.47 days. We identified the initial transit signal in KELT-North survey data. Archival and follow-up photometry, the Gaia parallax, radial velocities, Doppler tomography, and adaptive optics imaging were used to confirm the planetary nature of the companion and characterize the system. From global modeling we infer that the host star HD 185603 is a rapidly-rotating (VsinI~120 km/s) A2V star with an effective temperature of $T_{eff}$=8730K, mass of $M_{star}=1.76M_{sun}$, radius of $R_{star}=1.561R_{sun}$, surface gravity of logg=4.292, and age of <600 Myr. The planetary companion has a radius of $1.735^{+0.070}_{-0.075}~R_{J}$, a semimajor axis of $a=0.0542^{+0.0014}_{-0.0021}$AU, and a linear ephemeris of $BJD_{TDB}=2457503.120049 \pm 0.000190 + E(3.4741070\pm0.0000019)$. We place a $3\sigma$ upper limit of ~3.5 $M_{J}$ on the mass of the planet. The Doppler tomographic measurement indicates that the planetary orbit is well aligned with the projected spin-axis of the star ($\lambda= 3.4\pm {2.1}$ degrees). The inclination of the star is constrained to be $24.4<I_*<155.6$ degrees, implying a true (three-dimensional) spin-orbit alignment of $1.3<\psi<69.8$ degrees. The planet receives an insolation flux of $\sim 8\times 10^9~{\rm erg~s^{-1}~cm^{-2}}$, implying an equilibrium temperature of of ~ 2250 K, assuming zero albedo and complete heat redistribution. Due to the high stellar $T_{eff}$, the planet also receives an ultraviolet (wavelengths $d\le 91.2$~nm) insolation flux of $\sim 9.1\times 10^4~{\rm erg~s^{-1}~cm^{-2}}$, which may lead to significant ablation of the planetary atmosphere. Together with WASP-33, Kepler-13 A, HAT-P-57, KELT-17, and KELT-9, KELT-20 is the sixth A star host of a transiting giant planet, and the third-brightest host (in V) of a transiting planet.

astro-ph.EP↗

arXiv2017

Standard FITS template for simulated astrophysical scenes with the WFIRST coronagraph

The science investigation teams (SITs) for the WFIRST coronagraphic instrument have begun studying the capabilities of the instrument to directly image reflected light off from exoplanets at contrasts down to contrasts of ~10^-9 with respect to the stellar flux. Detection of point sources at these high contrasts requires yield estimates and detailed modeling of the image of the planetary system as it propagates through the telescope optics. While the SITs might generate custom astrophysical scenes, the integrated model, propagated through the internal speckle field, is typically done at JPL. In this white paper, we present a standard file format to ensure a single distribution system between those who produce the raw astrophysical scenes, and JPL modelers who incorporate those scenes into their optical modeling. At its core, our custom file format uses FITS files, and incorporates standards on packaging astrophysical scenes. This includes spectral and astrometric information for planetary and stellar point sources, zodiacal light and extragalactic sources that may appear as contaminants. Adhering to such a uniform data distribution format is necessary, as it ensures seamless work flow between the SITs and modelers at JPL for the goals of understanding limits of the WFIRST coronagraphic instrument.

astro-ph.IM↗

arXiv2016

The Orbit and Transit Prospects for $\beta$ Pictoris b constrained with One Milliarcsecond Astrometry

A principal scientific goal of the Gemini Planet Imager (GPI) is obtaining milliarcsecond astrometry to constrain exoplanet orbits. However, astrometry of directly imaged exoplanets is subject to biases, systematic errors, and speckle noise. Here we describe an analytical procedure to forward model the signal of an exoplanet that accounts for both the observing strategy (angular and spectral differential imaging) and the data reduction method (Karhunen-Lo\`eve Image Projection algorithm). We use this forward model to measure the position of an exoplanet in a Bayesian framework employing Gaussian processes and Markov chain Monte Carlo (MCMC) to account for correlated noise. In the case of GPI data on $\beta$ Pic b, this technique, which we call Bayesian KLIP-FM Astrometry (BKA), outperforms previous techniques and yields 1$\sigma$-errors at or below the one milliarcsecond level. We validate BKA by fitting a Keplerian orbit to twelve GPI observations along with previous astrometry from other instruments. The statistical properties of the residuals confirm that BKA is accurate and correctly estimates astrometric errors. Our constraints on the orbit of $\beta$ Pic b firmly rule out the possibility of a transit of the planet at 10-$\sigma$ significance. However, we confirm that the Hill sphere of $\beta$ Pic b will transit, giving us a rare chance to probe the circumplanetary environment of a young, evolving exoplanet. We provide an ephemeris for photometric monitoring of the Hill sphere transit event, which will begin at the start of April in 2017 and finish at the end of January in 2018.

astro-ph.EP↗

arXiv2016

The Peculiar Debris Disk of HD 111520 as Resolved by the Gemini Planet Imager

Using the Gemini Planet Imager (GPI), we have resolved the circumstellar debris disk around HD 111520 at a projected range of ~30-100 AU in both total and polarized $H$-band intensity. The disk is seen edge-on at a position angle of ~165$^{\circ}$ along the spine of emission. A slight inclination or asymmetric warping are covariant and alters the interpretation of the observed disk emission. We employ 3 point spread function (PSF) subtraction methods to reduce the stellar glare and instrumental artifacts to confirm that there is a roughly 2:1 brightness asymmetry between the NW and SE extension. This specific feature makes HD 111520 the most extreme examples of asymmetric debris disks observed in scattered light among similar highly inclined systems, such as HD 15115 and HD 106906. We further identify a tentative localized brightness enhancement and scale height enhancement associated with the disk at ~40 AU away from the star on the SE extension. We also find that the fractional polarization rises from 10 to 40% from 0.5" to 0.8" from the star. The combination of large brightness asymmetry and symmetric polarization fraction leads us to believe that an azimuthal dust density variation is causing the observed asymmetry.

astro-ph.EP↗