arXiv subjects

Onkar Jadhav

Publications and source records attributed to Onkar Jadhav.

5 recordsLinked to original sources

arXiv2026

Deep Learning-Based Statistical Downscaling of Sea Surface Temperature Using a Residual Corrective Neural Network

The large-scale oceanic and atmospheric forecasts provided by global climate models typically lack sufficient resolution to accurately capture the response of the coastal ocean to atmospheric forcing and coastal circulation that drive fine-scale SST variability. Dynamical downscaling is computationally prohibitive, when applied to extensive coastlines, predictive ensembles, or long time periods. Therefore, this work presents a statistical downscaling of sea surface temperature (SST) from the seasonal coupled ocean-atmosphere forecast system (ACCESS-S2) using machine learning techniques. This study proposes a novel deep learning framework that uses a U-Net to generate an initial high-resolution SST estimate, which is subsequently refined using a residual corrective approach. The target SST fields are derived from the Regional Ocean Modeling System (ROMS). This two step approach called Residual Corrective Neural Network (RCNN) progressively refines initial U-Net predictions by incorporating dynamically scaled residuals at each step, enabling accurate capture of broad patterns and fine-grained features such as eddies and fronts. We also introduce a custom loss-assisted RCNN variant to improve performance during extreme events, which may be absent from training data due to climate-driven shifts in SST extremes. The framework efficiently downscales SST along the west coast of Australia. A 2011 marine heatwave case study shows that the RCNN improves ACCESS-S2 SST predictions by increasing horizontal resolution from 25 km to 2 km, enabling identification of fine-scale anomalies unresolved in the ACCESS-S2 dataset. This balance between computational efficiency and accuracy supports applications in coastal impact assessment and marine ecosystem studies.

physics.ao-ph↗

arXiv2026

Patch-PODiff-ViT: Structured Latent Diffusion with Patchwise POD for Super-Resolution and Uncertainty Quantification

Diffusion models enable probabilistic super-resolution and conditional generation, but pixel-space methods are computationally expensive and learned latent spaces often lack interpretable uncertainty quantification. We introduce Patch-PODiff-ViT, a structured latent diffusion framework in which the latent space is defined by patchwise Proper Orthogonal Decomposition (POD), a fixed linear orthonormal basis over local patches, rather than learned by a nonlinear autoencoder. This yields low-dimensional, variance-ordered tokens that preserve spatial structure and enable efficient diffusion in a structured low-dimensional latent space with a Vision Transformer. Because the decoder is fixed, linear, and orthonormal, latent coefficient uncertainty can be propagated directly to physical-space predictive variance, enabling analytic propagation of predictive variance through the linear decoder without Monte Carlo estimation in pixel space. Across sea surface temperature, medical imaging, and natural images, the method achieves strong reconstruction with fewer parameters and lower memory, while producing well-calibrated spatial uncertainty that closely matches empirical ensembles.

cs.LG↗

arXiv2026

PODiff: Latent Diffusion in Proper Orthogonal Decomposition Space for Scientific Super-Resolution

Probabilistic super-resolution of high-dimensional spatial fields using diffusion models is often computationally prohibitive due to the cost of operating directly in pixel space. We propose PODiff, a structured conditional generative framework that performs diffusion in a fixed, variance-ordered Proper Orthogonal Decomposition (POD) coefficient space, exploiting the orthogonality of POD modes to impose an interpretable, variance-ordered latent geometry. This design enables efficient ensemble generation, preserves dominant spatial structure, and yields spatially interpretable, well-calibrated uncertainty at substantially lower computational cost. We evaluate PODiff on sea surface temperature downscaling over the West Australian coast and on a controlled advection-diffusion benchmark. PODiff achieves reconstruction accuracy comparable to pixel-space diffusion while requiring significantly less memory and producing more reliable uncertainty estimates than deterministic and Monte Carlo Dropout baselines.

cs.LG↗

arXiv2021

Error Analysis of a Model Order Reduction Framework for Financial Risk Analysis

A parametric model order reduction (MOR) approach for simulating the high dimensional models arising in financial risk analysis is proposed on the basis of the proper orthogonal decomposition (POD) approach to generate small model approximations for the high dimensional parametric convection-diffusion reaction partial differential equations (PDE). The proposed technique uses an adaptive greedy sampling approach based on surrogate modeling to efficiently locate the most relevant training parameters, thus generating the optimal reduced basis. The best suitable reduced model is procured such that the total error is less than a user-defined tolerance. The three major errors considered are the discretization error associated with the full model obtained by discretizing the PDE, the model order reduction error, and the parameter sampling error. The developed technique is analyzed, implemented, and tested on industrial data of a puttable steepener under the two-factor Hull-White model. The results illustrate that the reduced model provides a significant speedup with excellent accuracy over a full model approach, demonstrating its potential applications in the historical or Monte Carlo value at risk calculations.

math.NA↗

arXiv2020

Model order reduction for parametric high dimensional models in the analysis of financial risk

This paper presents a model order reduction (MOR) approach for high dimensional problems in the analysis of financial risk. To understand the financial risks and possible outcomes, we have to perform several thousand simulations of the underlying product. These simulations are expensive and create a need for efficient computational performance. Thus, to tackle this problem, we establish a MOR approach based on a proper orthogonal decomposition (POD) method. The study involves the computations of high dimensional parametric convection-diffusion reaction partial differential equations (PDEs). POD requires to solve the high dimensional model at some parameter values to generate a reduced-order basis. We propose an adaptive greedy sampling technique based on surrogate modeling for the selection of the sample parameter set that is analyzed, implemented, and tested on the industrial data. The results obtained for the numerical example of a floater with cap and floor under the Hull-White model indicate that the MOR approach works well for short-rate models.

q-fin.CP↗