Projects

Current Projects

• Biomedical Natural Language Processing (BioNLP)
  • Deep learning for biomedical text mining: Design of novel and application of existing cutting-edge NLP models (e.g. Transformer, BERT, and GPT-2) for biomedical information retrieval and text summarization. This is an unprecedented opportunity to harness unstructured information from millions of documents, use this information to guide generic drug review, and even drug discovery.
  • Using brain activity (e.g. fMRI) to interpret NLP models and guide model development.
• AI and NLP for Early Detection of Cognitive Decline
  • Alzheimer’s disease (AD) is a currently incurable brain disorder. Speech, a quintessentially human ability, has emerged as an important biomarker of neurodegenerative disorders like AD. Can AI-driven speech analysis help identify AD? In this project, we plan to test whether, or the extent to which, the natural language processing (NLP) model is capable of predicting dementia from spontaneous speech.
  • Media Coverage: [Drexel News][ScienceDaily][AAAS][AIMBE][Psychology Today][Interesting Engineering][Jerusalem Post][More Press Releases]
• Development of novel and application of existing deep generative models (e.g. GANs, VAEs) to explain the high-dimensional structure and time course of neural population activity. The focus is on the extraction of low-dimensional temporal patterns in high-dimensional spiking and local field potentials datasets in visual perception, visual attention and working memory tasks, and the development of new tools for causal inference (e.g. copula Granger causality).
• Deep learning for brain age prediction using multimodal neuroimaging data (structural MRI, DTI and resting-state fMRI)
• Topological Data Analysis (See our recent work on structure-function topological mapping and persistent homolgy)

Previous Projects

For details, visit our previous website.