Noise-Aware Visual Representation Learning for Medical Visual Question Answering
The framework improves robustness to noisy visual embeddings while maintaining competitive performance on clean inputs.
AI research + software
I research medical visual question answering and build software for people, institutions, and public services.
A denoising autoencoder learns to recover clean visual embeddings before they are projected into a language model.
The framework improves robustness to noisy visual embeddings while maintaining competitive performance on clean inputs.
Train with controlled noise in visual embeddings.
Recover clean features with a denoising autoencoder.
Map visual features into the language model space.
Adapt efficiently with LoRA for medical VQA.
Selected applications
Academic infrastructure, publishing, exam practice, and public data exploration.
Official program site for education, research, curriculum, and campus news.
An academic publishing hub for journals, books, and enduring ideas.
Self-directed CAT practice for TWK, TIU, and TKP with progress tracking.
Find public cameras across Bali and inspect available streams.
I work across machine learning and software engineering, moving from careful experiments to interfaces that make complex systems approachable.
Read the complete method, experiments, and findings on arXiv.
Read the research