Searcharxiv⌕ Search

arXiv subjects

Ashu Gupta

Publications and source records attributed to Ashu Gupta.

6 recordsLinked to original sources

TransXplainRRG: A Radiologist-Evaluated LVLM for Chest Radiograph Report Generation

Recent advancements in Large Vision-Language Models (LVLM) enhance X-ray report generation, but often neglect clinical context, overlook workflow integration, and lack explainability, which hinders clinical adoption due to interpretability and reliability concerns. This paper addresses the dual challenges of performance and expert explainability in automated X-ray report generation for clinical AI applications. Our proposed approach, TransXplainRRG, utilises an off-the-shelf Swin Transformer model along with a transformer-based text encoder that incorporates patient medical history to generate a radiology report. Further, we explore an expert-guided `inside-out' approach and extract only abnormal findings for radiology report refinement. Thus, this study bridges the gap between high-performance automation and interpretability critical for clinical practice by combining state-of-the-art transformer-based vision encoders, text encoders, and LLMs. TransXplainRRG is trained and tested on the large-scale MIMIC-CXR dataset and further evaluated on the unseen IU X-Ray dataset to demonstrate its generalizability and robustness. We employed the GREEN metric as a complementary measure of clinical consistency, achieving scores of 0.327 on MIMIC-CXR and 0.605 on IU X-Ray, both higher than those achieved by MAIRA-2 under the same evaluation setting. It also performed competitively across standard evaluation metrics, frequently matching or surpassing state-of-the-art models. Additionally, we introduce qualitative evaluation metrics developed from radiologists' viewpoints and conduct a preliminary radiologist evaluation on 100 generated reports to assess their clinical relevance. Quantitative and qualitative analyses further demonstrate the potential of our TransXplainRRG framework for reliable and explainable radiology report generation.

cs.CV↗

A Needs Assessment for Measuring Geographic - Legislative Associations in the U.S. House of Representatives

Political legislation affects the well-being and livelihoods of constituents. In the U.S. Congress a representative's voting record on bills and legislation is public. These bills have themes associated with them, such as veterans' affairs, coastal monitoring, agricultural appropriations, etc. A bill on veterans' affairs may affect a constituency differently if they have a high percentage of veterans. In this work, we demonstrate how congressional vote outcomes can be merged with typical geographic information systems (GIS) data to help compare a legislator's votes with the geographies of their constituencies to measure the association between district features and legislators' decisions. We retrieved and tagged bills from the 118th U.S. House of Representatives (Jan. 2023 - Jan. 2025) by manually assigning each bill a set of themes. We then retrieved spatial data at the congressional district level related to each theme. We built a backend database to connect bill information and spatial data, and an interactive map prototype displayed the spatial data related to each theme associated with each bill. Our work helps identify challenges to creating a more accessible and complete system that would facilitate the visualization and analysis of the nature of contemporary Congressional representation. This technical report is primarily written for members of the data science community, and specifically those interested in legislative and geographic data.

cs.HC↗

A Utility-preserving De-identification Pipeline for Cross-hospital Radiology Data Sharing

Large-scale radiology data are critical for developing robust medical AI systems. However, sharing such data across hospitals remains heavily constrained by privacy concerns. Existing de-identification research in radiology mainly focus on removing identifiable information to enable compliant data release. Yet whether de-identified radiology data can still preserve sufficient utility for large-scale vision-language model training and cross-hospital transfer remains underexplored. In this paper, we introduce a utility-preserving de-identification pipeline (UPDP) for cross-hospital radiology data sharing. Specifically, we compile a blacklist of privacy-sensitive terms and a whitelist of pathology-related terms. For radiology images, we use a generative filtering mechanism that synthesis a privacy-filtered and pathology-reserved counterparts of the original images. These synthetic image counterparts, together with ID-filtered reports, can then be securely shared across hospitals for downstream model development and evaluation. Experiments on public chest X-ray benchmarks demonstrate that our method effectively removes privacy-sensitive information while preserving diagnostically relevant pathology cues. Models trained on the de-identified data maintain competitive diagnostic accuracy compared with those trained on the original data, while exhibiting a marked decline in identity-related accuracy, confirming effective privacy protection. In the cross-hospital setting, we further show that de-identified data can be combined with local data to yield better performance.

cs.CV↗

Improving Gene Trees without more data

Estimating species and gene trees from sequence data is challenging. Gene tree estimation is often hampered by low phylogenetic signal in alignments, leading to inaccurate trees. Species tree estimation is complicated by incomplete lineage sorting (ILS), where gene histories differ from the species' history. Summary methods like MP-EST, ASTRAL2, and ASTRID infer species trees from gene trees but suffer when gene tree accuracy is low. To address this, the Statistical Binning (SB) and Weighted Statistical Binning (WSB) pipelines were developed to improve gene tree estimation. However, previous studies only tested these pipelines using multi-locus bootstrapping (MLBS), not the BestML approach. This thesis proposes a novel pipeline, WSB+WQMC, which shares design features with the existing WSB+CAML pipeline but has other desirable properties and is statistically consistent under the GTR+MSC model. This study evaluated WSB+WQMC against WSB+CAML using BestML analysis on various simulated datasets. The results confirmed many trends seen in prior MLBS analyses. WSB+WQMC substantially improved gene tree and species tree accuracy (using ASTRAL2 and ASTRID) on most datasets with low, medium, and moderately high ILS levels. In a direct comparison, WSB+WQMC computed less accurate trees than WSB+CAML under certain low and medium ILS conditions. However, WSB+WQMC performed better or at least as accurately as WSB+CAML on all datasets with moderately high and high ILS. It also proved better for estimating gene trees on some medium and low ILS datasets. Thus, WSB+WQMC is a promising alternative to WSB+CAML for phylogenetic estimation, especially in the presence of low phylogenetic signal.

q-bio.PE↗

Unsupervised Search for Ethnic Minorities' Medical Segmentation Training Set

This article investigates the critical issue of dataset bias in medical imaging, with a particular emphasis on racial disparities caused by uneven population distribution in dataset collection. Our analysis reveals that medical segmentation datasets are significantly biased, primarily influenced by the demographic composition of their collection sites. For instance, Scanning Laser Ophthalmoscopy (SLO) fundus datasets collected in the United States predominantly feature images of White individuals, with minority racial groups underrepresented. This imbalance can result in biased model performance and inequitable clinical outcomes, particularly for minority populations. To address this challenge, we propose a novel training set search strategy aimed at reducing these biases by focusing on underrepresented racial groups. Our approach utilizes existing datasets and employs a simple greedy algorithm to identify source images that closely match the target domain distribution. By selecting training data that aligns more closely with the characteristics of minority populations, our strategy improves the accuracy of medical segmentation models on specific minorities, i.e., Black. Our experimental results demonstrate the effectiveness of this approach in mitigating bias. We also discuss the broader societal implications, highlighting how addressing these disparities can contribute to more equitable healthcare outcomes.

cs.CV↗

Derandomizing Isolation Lemma for $K_{3,3}$-free and $K_5$-free Bipartite Graphs

The perfect matching problem has a randomized NC algorithm, using the celebrated Isolation Lemma of Mulmuley, Vazirani and Vazirani. The Isolation Lemma states that giving a random weight assignment to the edges of a graph, ensures that it has a unique minimum weight perfect matching, with a good probability. We derandomize this lemma for $K_{3,3}$-free and $K_5$-free bipartite graphs, i.e. we give a deterministic log-space construction of such a weight assignment for these graphs. Such a construction was known previously for planar bipartite graphs. Our result implies that the perfect matching problem for $K_{3,3}$-free and $K_5$-free bipartite graphs is in SPL. It also gives an alternate proof for an already known result -- reachability for $K_{3,3}$-free and $K_5$-free graphs is in UL.

cs.CC↗