Tag
This paper proposes HyMLRaman, a hybrid framework combining EfficientNet-B3 spectral embeddings with classical classifiers (notably SVM) and DDPM-based generative feature augmentation to identify six pharmaceutical compounds via Raman spectroscopy, achieving 96.31% accuracy.
This paper demonstrates the use of Sharpness-Aware Minimization (SAM) to enhance the generalization of machine learning models for classifying bacterial Raman spectral data, improving accuracy by up to 10.5% and advancing portable diagnostic tools for antimicrobial resistance.
The study establishes a physics-informed AI framework using Raman spectroscopy and machine learning to authenticate edible oils, achieving high classification accuracy with compact spectral representations.
This paper proposes a Gated Bottleneck Latent ODE combined with Multi-Path Just-In-Time Fine-Tuning and Raman data fusion to improve multi-day forecasting of mammalian cell culture processes, achieving better performance on real bioreactor data.
Researchers at MIT have developed PlasmoSniff, a breath test that uses nanoparticles and enhanced Raman spectroscopy to detect pneumonia and other lung conditions in minutes, potentially enabling rapid at-home or clinical diagnosis.
This review examines the role of machine learning across the biomedical Raman spectroscopy pipeline, from preprocessing and signal correction to clinical translation, highlighting barriers and future directions for reliable Raman-AI systems.