Portfolio item number 1
Short description of portfolio item number 1
Short description of portfolio item number 1
Short description of portfolio item number 2 
Published in Indian Patent Application No.: 202521024259, 2025
Patent application on projecting synthetic nodules in medical imaging.
Recommended citation: Goyal, A. (2025). "Patent: A System and Method for Projecting Synthetic Nodules in Medical Imaging". Indian Patent Application No.: 202521024259
Published in DEMI Workshop at MICCAI, 2025
A diffusion model that turns blurry chest X-rays projected from CT scans into sharp, realistic X-rays while keeping small lung nodules visible.
Recommended citation: Goyal, A.*, Mittal, A.*, Rao, P., Tadepalli, M., and Putha, P. (2025). "DiffusionXRay: A Diffusion and GAN-Based Approach for Enhancing Digitally Reconstructed Chest Radiographs". DEMI Workshop at MICCAI. (*equal contribution)
Download Paper
Published in B.Tech-M.Tech Dual Degree Project • IIT Bombay, 2025
Forecasting MSW generation using economic indicators and estimating waste volumes from single RGB images using monocular depth inference.
Recommended citation: Goyal, A. (2025). "Forecasting and Volume Estimation of Municipal Solid Waste Using Machine Learning". IIT Bombay.
Published in Medical Imaging with Deep Learning (MIDL), 2026
A diffusion model that generates realistic lung nodules on chest X-rays with control over border, texture, calcification and subtlety, and uses them to train better nodule detectors.
Recommended citation: Goyal, A.*, Singh, S.*, Mittal, A.*, Tadepalli, M., Kumar, P., and Putha, P. (2026). "A Diffusion-Driven Fine-Grained Nodule Synthesis Framework for Enhanced Lung Nodule Detection from Chest Radiographs". Medical Imaging with Deep Learning (MIDL). (*equal contribution)
Download Paper
Published in Mechanistic Interpretability Workshop at ICML, 2026
A position paper on which LLM interpretability tools still work on vision-language-action models, which silently break, and why.
Recommended citation: Goyal, A. (2026). "The Hitchhiker's Guide to Mechanistic Interpretability of Vision-Language-Action Models". Mechanistic Interpretability Workshop at ICML.
Download Paper
arXiv preprint, 2026
Per-state, finite-horizon measurements of whether a small action error grows or dies out under open-loop execution and closed-loop replanning, across twelve manipulation tasks from robomimic, MimicGen and RoboCasa.
Recommended citation: Goyal, A. (2026). "Measuring the Stability Assumption Behind Action Chunking". arXiv:2610.01626.
Download Paper