A5008387529- Machine Learning in Healthcare
- Advanced Graph Neural Networks
- Radiomics and Machine Learning in Medical Imaging
- Artificial Intelligence in Healthcare
- AI in cancer detection
- Functional Brain Connectivity Studies
- Brain Tumor Detection and Classification
- Health, Environment, Cognitive Aging
- Artificial Intelligence in Healthcare and Education
- Advanced Neuroimaging Techniques and Applications
- Bioinformatics and Genomic Networks
- Topic Modeling
- Image Retrieval and Classification Techniques
- Medical Imaging and Analysis
- Dementia and Cognitive Impairment Research
- Medical Imaging Techniques and Applications
- COVID-19 diagnosis using AI
- Imbalanced Data Classification Techniques
- Advanced Image and Video Retrieval Techniques
- Multimodal Machine Learning Applications
- Computational Drug Discovery Methods
- Neural Networks and Applications
- Hip and Femur Fractures
- Text and Document Classification Technologies
- Medical Image Segmentation Techniques
Massachusetts General Hospital
2023-2025
Harvard University
2023-2025
Boston University
2024
Athinoula A. Martinos Center for Biomedical Imaging
2023-2024
Technical University of Munich
2016-2023
Harvard University Press
2022
Indian Institute of Technology Kharagpur
2016-2018
Graph deep learning has recently emerged as a powerful ML concept allowing to generalize successful neural architectures non-euclidean structured data. Such methods have shown promising results on broad spectrum of applications ranging from social science, biomedicine, and particle physics computer vision, graphics, chemistry. One the limitations majority current graph network is that they are often restricted transductive setting rely assumption underlying known fixed. Often, this not true...
Abstract The standard of care for brain tumors is maximal safe surgical resection. Neuronavigation augments the surgeon’s ability to achieve this but loses validity as surgery progresses due shift. Moreover, gliomas are often indistinguishable from surrounding healthy tissue. Intraoperative magnetic resonance imaging (iMRI) and ultrasound (iUS) help visualize tumor iUS faster easier incorporate into workflows offers a lower contrast between tumorous tissues than iMRI. With success...
Multi-modal data comprising imaging (MRI, fMRI, PET, etc.) and non-imaging (clinical test, demographics, can be collected together used for disease prediction. Such diverse gives complementary information about the patient's condition to make an informed diagnosis. A model capable of leveraging individuality each multi-modal is required better We propose a graph convolution based deep which takes into account distinctiveness element data. incorporate novel self-attention layer, weights every...
We tackle the prediction of age and mini-mental state examination (MMSE) score based on structural brain connectivity derived from diffusion magnetic resonance images. propose a machine-learning model inspired by graph convolutional networks (GCNs), which takes input processes data separately through parallel GCN mechanism with multiple branches, thereby disentangling node features. The novelty our work lies in architecture, especially attention module, learns an embedding representation...
The standard of care for brain tumors is maximal safe surgical resection. Neuronavigation augments the surgeon's ability to achieve this but loses validity as surgery progresses due shift. Moreover, gliomas are often indistinguishable from surrounding healthy tissue. Intraoperative magnetic resonance imaging (iMRI) and ultrasound (iUS) help visualize tumor iUS faster easier incorporate into workflows offers a lower contrast between tumorous tissues than iMRI. With success data-hungry...
Interpretability in Graph Convolutional Networks (GCNs) has been explored to some extent computer vision general, yet, the medical domain, it requires further examination. Moreover, most of interpretability approaches for GCNs, especially focus on interpreting model a post hoc fashion. In this paper, we propose an interpretable graph learning-based which 1) interprets clinical relevance input features towards task, 2) uses explanation improve performance and, 3) learns population level...