KR Mangalam University

Human–resource (HR) models often fail when deployed across heterogeneous organizational units because correlations are confounded by selection and degrade under environment shift. We present CIVIC-IT, a Causal–Invariant–Conformal framework that jointly targets identification, transfer, and calibrated uncertainty for HR decision making. First, we estimate average and heterogeneous effects of environment-level diversity levers using cross-fitted double/debiased machine learning with cluster-robust inference. Second, we learn shift-robust representa- tions via invariant risk minimization (IRM) combined with a domain-adversarial objective (DANN) and fine-tune with Group–DRO to control worst-environment risk. Third, we attach distribution-free guarantees using split-conformal predic- tion for continuous outcomes and label-conditional conformal predictive sets for a “high T ∗” (turnover intention) flag. Outcomes are psychometrically grounded latent scores for Inclusion, Engagement, and T ∗ built from ordinal indicators with reliability weighting and multi-group diagnostics for approximate invari- ance. Using the IBM HR Attrition dataset (N =1,470) configured into realistic environments (Department, JobRole, and Department×Role), we find: (i) global ATEs are small, but policy-relevant heterogeneity emerges—e.g., for early- career Research Scientist teams, greater diversity of education field increases Inclusion/Engagement, whereas highly heterogeneous business–travel patterns reduce them; (ii) leave-one-environment-out evaluation identifies stable trans- fer with clear best/worst environments and improved worst-group error under Group–DRO; (iii) conformal intervals achieve near-nominal coverage, and the risk classifier yields predominantly singleton predictive sets at the target error rate; and (iv) multi-objective policy search produces fairness-aware Pareto port- folios that deliver concurrent (albeit modest) gains in Inclusion and Engagement with slight reductions in T ∗. CIVIC-IToffers a governance-ready blueprint that links robust estimation to auditable, deployable HR policy in heterogeneous IT environments.

Image dehazing plays a critical role in understanding complex real-world systems such as transportation, environmental monitoring, and healthcare imaging, where haze significantly degrades visual information and hinders decision-making. Traditional deep learning models emphasize accuracy but often neglect interpretability and resource efficiency, limiting their applicability in sustainable multi-modal fusion frameworks. We introduce SFX-GAN, a Sustainable and eXplainable Frequency-domain GAN that employs a novel channel-wise spectral fusion strategy, treating low, mid, and high-frequency components across RGB channels as complementary modalities. This multi-modal spectral integration enables the model to capture both global context and fine-grained structural cues, facilitating more reliable restoration of haze-free images. For interpretability, SFX-GAN incorporates a Concept Bottleneck Layer and Grad-CAM supervision, aligning network attention with haze-affected regions and enhancing trust in automated dehazing for safety-critical applications. Designed with sustainability in mind, SFX-GAN leverages frequency-domain processing to deliver efficient inference with reduced computational overhead. Experiments on RESIDE-SOTS, NH-HAZE, I-HAZE, O-HAZE, and D-HAZY datasets demonstrate state-of-the-art performance under both full-reference (PSNR, SSIM) and no-reference (NIQE, BRISQUE, FADE, SSEQ) metrics. Ablation studies confirm the effectiveness of each module in the fusion pipeline. Overall, SFX-GAN contributes to the intelligent fusion of spectral-spatial modalities for complex imaging systems, advancing responsible and explainable AI aligned with SDG 9 (Industry, Innovation, and Infrastructure) and SDG 11 (Sustainable Cities and Communities).

Diabetic retinopathy (DR) is a gradual and severe consequence of diabetes that increases the risk of visual impairment and blindness globally. Early and precise identification of DR is crucial for successful care, but availability to specialist diagnostic tools is restricted, especially in low-resource settings. This study aims to improve the automated identification of diabetic retinopathy using cutting-edge deep learning models—specifically ViT, CLIP, BLIP, ViLT, Florence-2, and ResNet50—which are applied to the APTOS Diabetic Retinopathy dataset. Diabetic Retinopathy (DR) diagnosis is often hindered by poor image quality and subtle differences across varying severity levels. The grading of DR is typically categorized into five stages: No DR, Mild, Moderate, Severe, and Proliferative, as defined by the International Clinical Diabetic Retinopathy Disease Severity Scale. To address artifacts such as uneven illumination, we employ Gaussian preprocessing to smooth noise and Wavelet transforms to enhance edge details. This study evaluates the impact of Gaussian and Wavelet preprocessing techniques on the performance of Vision-Language Models (e.g., CLIP, BLIP) and CNNs (e.g., RESNET50). Results indicate that Gaussian Filtering significantly enhances classification accuracy, yielding a 37% improvement for RESNET50 and 36% for BLIP compared to the baseline. This study demonstrates that appropriate preprocessing is critical for maximizing the potential of deep learning in medical imaging. Specifically, Gaussian Filtering proved superior to Wavelet techniques, offering substantial accuracy gains for models like BLIP and RESNET50. Future work will explore hybrid preprocessing pipelines to further distinguish between the subtle early stages of the disease.

The paper examines the impact of firm size and industry type on Environmental, Social and Governance (ESG) performance of 262 listed Indian companies between 2018 and 2022, with a specific focus on an emerging-market context. Using Refinitiv ESG scores, the study employs Panel-Corrected Standard Errors (PCSE) regression to address heteroskedasticity, serial correlation, and cross-sectional dependence in longitudinal ESG data, thereby enhancing the robustness of empirical inference. The findings indicate that larger firms exhibit superior ESG performance due to greater resource availability and stakeholder visibility, while service-sector firms outperform manufacturing firms owing to lower environmental impact and higher transparency. However, the sectoral gap in ESG performance narrows following the adoption of SEBI’s Business Responsibility and Sustainability Reporting (BRSR) framework, highlighting the mediating role of regulatory intervention in shaping corporate sustainability behavior. The study contributes to ESG literature by combining methodological rigor through PCSE estimation with institutional analysis of the Indian regulatory environment, and by extending legitimacy theory to explain how firm characteristics and regulatory mandates jointly influence ESG performance in an emerging economy. The findings offer valuable insights for regulators, managers, and investors seeking to strengthen sustainable corporate governance amid evolving disclosure regimes.