My work focuses on developing and applying machine learning and computational methods to scientific problems.
I’m also interested in language models and broader problems in modern machine learning.
Feel free to reach out if you’d like to discuss an interesting project or research idea.
Uses PCA-reduced InSAR histories and a random forest to forecast landslide instability seven days ahead, asking whether noisy satellite deformation can support short-horizon prediction rather than retrospective detection.
The undeformed glass contains enough structural information to predict where its shear band will form. Softening a targeted region before loading can then move the eventual band to a different location.
Weak-form constraints transfer derivatives from noisy data to smooth test functions, stabilizing inverse parameter recovery in scientific imaging problems where pointwise physics-informed losses amplify noise.
Combines PCA-reduced InSAR deformation histories with a random forest to forecast instability seven days ahead. The study asks whether satellite displacement signals can support short-horizon prediction, rather than only retrospective detection.
Uses 2D CNN pretraining to give a 3D CT model a stronger starting point when labeled volumetric data are scarce. The method was developed for COVID-19 diagnosis, where full 3D training sets are relatively small.
Manuscripts currently under peer review.
Weak-form constraints transfer derivatives from noisy data to smooth test functions, stabilizing inverse parameter recovery on scientific imaging problems where pointwise physics-informed losses amplify noise.
Internal activations add evidence when deceptive or collusive information is hidden from the transcript. When visible behavior already reveals it, white-box probes add much less beyond the behavioral evidence.
Small stabilization changes can leave heat-transfer outputs nearly unchanged even when the solver has changed. Modified-equation signatures reveal those hidden numerical changes directly from the computed solution fields.
A localized perturbation spreads through a structural glass inside a deterministic butterfly cone. Its speed changes little as structural relaxation slows dramatically, separating information spreading from glassy dynamics.
The undeformed glass contains enough structural information to predict where its shear band will form. Softening a targeted region before loading can then move the eventual band to a different location.
Machine Learning Intern
Built evaluations and benchmarks for frontier AI models.
Peer Reviewer
Academic service · Springer Nature
Research Collaborator
Medical imaging and machine learning research.
Researcher (Volunteer)
Data analysis and research.
Email: arnavgarg888@gmail.com