Scholar
Soha Hassoun
Google Scholar ID: tR5MNfkAAAAJ
Professor & Past Chair. Department of Computer Science, Tufts University
Citations & Impact
Citations
875
H-index
18
i10-index
23
Publications
20
Co-authors
30
list available
Publications
Resume
Academic Achievements
- OMG: De Novo molecular generation from spectra via transfer learning and curriculum learning.
- MADGEN: Generating de novo molecules from partial structures, guided by spectra.
- A novel paradigm for metabolite annotation: Avoiding explicit construction of spectra and molecules, instead comparing query spectrum against candidates in the embedding space.
- MassSpecGym: A community effort to standardize the evaluation of mass spec annotation techniques.
- Ensemble Spectral Prediction (ESP): A pipeline that includes molecular representation learning, spectral prediction with peak co-dependency analysis, and rank-based learning, improving performance by up to 41% over MLPs.
- Separate normalization of normal and [CLS] tokens in self-supervised transformers, showing a 2.7% performance improvement in image, natural, and graph tasks.
- Predicting enzyme-substrate interactions using contrastive multiview coding (CMC).
- Review article on recent advances in computational methods for metabolomics.
Research Experience
- Current professor at Tufts University, leading the Hassoun Lab, which focuses on developing machine learning and AI models, especially those tailored for biological data.
Education
- Tufts University
Background
- Professor of Computer Science (primary), Chemical & Biological Engineering, and Electrical & Computer Engineering. Research interests include machine learning and systems biology, focusing on developing analysis and design tools to advance biotechnology.
Miscellany
- Contact Info: 177 College Ave, Medford, MA 02155, soha (at) cs.tufts.edu, Follow @sohahassoun
