Research & Publications

GeoAI-Powered Traffic Modeling: A Framework for Predictive Spatial Analytics

Authors: A. Rozhkov, P. Motarwar, R. Patil.

Venue: Association of Collegiate Schools of Planning (ACSP) Annual Conference

Year: 2025

Proposed a novel GeoAI framework integrating spatial grids, LSTM forecasting, and multimodal data fusion for urban congestion prediction and interactive policymaking.

Hybrid Approach for Quantifying Company Assets Using Structural Credit Risk Models

Authors: P. Motarwar, R. Patil, A. S. Lembhe, R. Selvamani, A. Bhuvaneswari.

Venue: Sustainable Development in AI, Blockchain, and E-Governance Applications (Springer)

Year: 2023

Formulated a hybrid AI–finance model combining credit risk estimation and asset value prediction using firm-level indicators and stochastic modeling principles.

Enhancement in Skin Cancer Detection Using Image Super Resolution and Convolutional Neural Network

Authors: A. Lembhe, P. Motarwar, R. Patil, S. Elias.

Venue: Procedia Computer Science(vol. 218, pp. 164–173)

Year: 2023

Introduced an image super-resolution preprocessing pipeline coupled with CNN-based classification to enhance dermoscopic image quality and improve melanoma detection accuracy across low-resolution clinical datasets.

Cluster-Centric Based Hybrid Approach for Cricket Sports Analytics Using Machine Learning

Authors: R. Patil, A. Duraphe, P. Motarwar, G. Suganya, M. Premalatha.

Venue: International Conference on Deep Sciences for Computing and Communications

Year: 2022

Developed a hybrid machine learning pipeline combining clustering and regression methods to analyze team performance and player strategy metrics, contributing to data-driven sports analytics.

Cognitive Approach for Heart Disease Prediction Using Machine Learning

Authors: P. Motarwar, A. Duraphe, G. Suganya, M. Premalatha.

Venue: International Conference on Emerging Trends in Information Technology

Year: 2020

This paper presents a predictive framework leveraging supervised learning algorithms to identify cardiovascular risk using clinical indicators, achieving significant accuracy improvements through optimized feature selection.

Ongoing Research

Research Assistant, NYU Visualization and Data Analytics Research Center (VIDA)
Advisor: Prof. Dr. Robert Krueger | Sep 2024 – Present

Maintaining an open-source Python–Flask application enabling semi-automatic gating in multiplex immunofluorescence (IF) imaging. The tool combines computational scalability with interactive visualization, empowering biomedical researchers to refine Gaussian mixture model–based auto-gating through visual feedback.

LLM-Augmented Interpretation Layer for Visual Analytics: Designing an LLM-based explanation module to generate natural language summaries of spatial and cellular phenomena within 2D and 3D biomedical images. Exploring multimodal prompt strategies (zero-shot, few-shot, and fine-tuned) tailored for cancer imaging and spatial expression analysis.