Publications

Antibiotic resistance & genomic ML

• 2026

  FARM: Forecasting Antibiotic Resistance in Mycobacterium tuberculosis Using Biophysics and Machine Learning

  PreprintbioRxiv (2026)Under review at PNAS
  Summary

  Mahbuba Tasmin, Shrishti Barethiya, Yu Wang, Lulu Kang, Jianhan Chen, Anna G. Green

  Fuses protein thermostability and biophysical energetics with machine learning to forecast mutation-level antibiotic resistance under data scarcity.

• 2026

  BIG-TB: A Benchmark for Prediction and Interpretability of Sequence-Based Machine Learning Using Mycobacterium tuberculosis Genomes

  PreprintbioRxiv (2026)Under revision following peer review
  Summary

  Mahbuba Tasmin, Saishardha Mohanty, S. Kulkarni, Maha R. Farhat, Anna G. Green

  A unified 17K-isolate benchmark dataset for genotype-to-phenotype prediction across 11 WHO-priority antibiotics, integrating genomic, proteomic, and evolutionary modalities.

• 2025

  The Structural Context of Mutations in Proteins Predicts Their Effect on Antibiotic Resistance

  Journal articleeLife 14:RP109450 (2025)
  Summary

  Anna G. Green, Mahbuba Tasmin, Rodrigo Vargas Jr., Maha R. Farhat

  Protein structural context features yield state-of-the-art accuracy and interpretability for antibiotic resistance mutation prediction.

• 2025

  Beyond Sequence-Only Models: Leveraging Structural Constraints for Antibiotic Resistance Prediction in Sparse Genomic Datasets

  Workshop paperICLR 2025 MLGenX Workshop
  Summary

  Mahbuba Tasmin, Anna Green

  Structural constraints paired with deep learning improve resistance prediction under extreme label sparsity.

• 2024

  Protein Structure-Informed Regularized Linear Model Outperforms ESM for Predicting Antibiotic Resistance

  ConferenceProgram in Quantitative Genomics Conference (PQG), Harvard University
  Summary

  Mahbuba Tasmin

  Poster highlighting a fused regularized linear model that integrates 3D structural features and surpasses ESM-based baselines for resistance prediction.

Language models & medical AI

• 2025

  Unveiling GPT-4V’s Hidden Challenges Behind High Accuracy on USMLE Questions

  Journal articleJournal of Medical Internet Research (2025)
  Summary

  Zhichao Yang, Zonghai Yao, Mahbuba Tasmin, Parth Vashisht, Won Seok Jang, Feiyun Ouyang, Beining Wang, David McManus, Dan Berlowitz, Hong Yu

  Analyzes GPT-4V performance on medical licensing questions, revealing systematic failure modes masked by headline accuracy.

Earlier work in sensing & modeling

• 2022

  Assessment of Deep Learning Models for Human Activity Recognition on Multi-Variate Time Series Data and Non-Targeted Adversarial Attack

  Book chapterAdvances in Intelligent Systems Research and Innovation, Studies in Systems, Decision and Control, vol. 379, Springer, pp. 129–159 (2022)
  Summary

  Mahbuba Tasmin, Sharif Uddin Ruman, Taoseef Ishtiak, Arif-ur-Rahman Chowdhury Suhan, Redwan Hasif, Shahnawaz Zulminan, Rashedur M. Rahman

  Benchmarks deep learning classifiers for human activity recognition and studies the vulnerability of the resulting models to non-targeted adversarial attacks.

• 2021

  Non-Newtonian Effect on Heat Transfer and Entropy Generation of Natural Convection Nanofluid Flow Inside a Vertical Wavy Porous Cavity

  Journal articleSN Applied Sciences, 3:299 (2021)
  Summary

  Mahbuba Tasmin, Preetom Nag, Zarin T. Hoque, Md. Mamun Molla

  A numerical study of heat transfer and entropy generation for non-Newtonian nanofluid flow in a differentially heated, wavy porous cavity.

• 2020

  Comparative Study of Classifiers on Human Activity Recognition by Different Feature Engineering Techniques

  ConferenceProceedings of the 2020 IEEE 10th International Conference on Intelligent Systems, pp. 93–101
  Summary

  Mahbuba Tasmin, Sifat Jahan, Taoseef Ishtiak, Sajid Ahmed, Sharif Uddin Ruman, Md. Shahnawaz Zulminan, Arif Ur Rahaman Chowdhury Suhan, Abdur Raufus Saleheen, N.M. Shihab Islam, Rashedur M. Rahman

  Compares five classifier models across feature sets produced by four feature-selection techniques for human activity recognition from ambient sensor data.

• 2018

  Multi-Dimensional Aspect Analysis of Text Input through Human Emotion and Social Factors

  Extended abstractACM International Joint Conference and Symposium on Pervasive and Ubiquitous Computing (UbiComp/ISWC '18 Adjunct), pp. 1779–1781, Singapore
  Summary

  Mahbuba Tasmin

  Proposes a multi-dimensional, aspect-based sentiment analysis capturing the influence of human emotion and social factors on text-input polarity.