معرفی
Marzyeh Ghassemi is an Associate Professor at MIT in the Department of Electrical Engineering and Computer Science (EECS) and the Institute for Medical Engineering & Science (IMES). She also holds a Canadian CIFAR AI Chair and is a faculty member at the Vector Institute. Her affiliations include the Jameel Clinic and CSAIL. Dr. Ghassemi's work focuses on ethical AI, algorithmic fairness in healthcare, and mitigating bias in medical datasets. She has pioneered projects analyzing racial disparities in end-of-life care, vocal disorder diagnosis using accelerometers, and improving subgroup robustness in machine learning.
Education:
- PhD in Computer Science, MIT
- MSc in Biomedical Engineering, University of Oxford (Marshall Scholar)
- B.S. in Computer Science and Electrical Engineering, New Mexico State University (Goldwater Scholar)
Research Interests: Her work bridges AI and healthcare, emphasizing fairness, reproducibility, and ethical deployment. Key areas include:
- Algorithmic Fairness: Addressing biases in medical imaging and clinical decision systems.
- Bias Mitigation: Developing methods to decorrelate sensitive information from predictions.
- Health Equity: Studying disparities in end-of-life care and organ procurement.
- Generative AI: Applications in mental health and medical imaging.
Publications: Her recent work highlights challenges in vision-language models (e.g., handling negation), regulatory gaps in health AI, and bias reduction techniques. These efforts aim to ensure AI systems are both accurate and equitable in clinical settings.
Awards:
- Herman L. F. von Helmholtz Career Development Professorship
- CIFAR Azrieli Global Scholar
- MIT Tech Review’s 35 Innovators Under 35
- Canada Research Chair
Leadership & Impact: She co-chaired the NeurIPS Workshop and founded the Association for Health Learning and Inference. Her projects like Quantifying Racial Disparities in End-of-Life Care reveal systemic biases in ICU treatments. She advocates for transparent AI deployment through frameworks like FUTURE-AI and TRIPOD+.
Labs & Teams: Leads MIT’s HealthyML group, collaborating on projects such as voice misuse detection, federated learning for medical imaging, and AI-driven mental health prediction.
