Research

I am a fourth-year Ph.D. candidate at Tufts University, where I am a part of the High-dimensional learning and data analysis (HILDA) research group. My Ph.D. advisor is James Murphy.

My thesis research analyzes multiscale structure in data. Typically, any single dataset contains multiple latent separations that could be considered correct. In one problem setting, a coarse separation of one’s dataset may be desired, while another problem setting may call for a finer separation within the data. In my thesis research, I leverage tools from harmonic analysis, graph theory, and linear algebra to build fast machine learning algorithms for finding and comparing these different latent separations within data.

I also am interested in medical signal processing. I have an affiliation with the Neuroanesthesia Research Laboratory at Massachusetts General Hospital under Dr. Oluwaseun Johnson-Akeju. In Dr. Akeju’s laboratory, I use machine learning algorithms on EEG and ECG data to predict brain states of patients receiving anesthetics.

Current projects:

  1. Applications of diffusion geometry to multiscale clustering problems.
  2. The inverse problem to the Affine Wealth Model: given that we want some distribution of wealth, what redistribution policy should we implement? Whom should we tax if we wish to decrease wealth inequality?

Submitted Papers:

  1. Sam L. Polk and Bruce M. Boghosian. (2019) “The Non-Universality of Wealth Distribution Tails Near Wealth Condensation Criticality.” Submitted to SIAM Journal of Applied Mathematics. (Link)
  2. Sam L. Polk*, Kimia Kashkooli*, …, and Oluwaseun Akeju (2020) “ECG-Derived Autonomic Nervous System Dynamics Predict Anaesthetic states.” Submitted to Biomedical Signal Processing and Control.

Published Papers:

  1. Kimia K. Kashkooli*, Sam L. Polk*, …, Oluwaseun Akeju, and Shubham Chamadia (2020) “Improved Tracking of Sevoflurane Anesthetic States with Drug-Specific Machine Learning Models.” Journal of Neural Engineering. (Link)
  2. Sam L. Polk*, Kimia Kashkooli*, …, and Oluwaseun Akeju. (2019). “Automatic Detection of General Anesthetic-States using ECG-Derived Autonomic Nervous System Features.” 2019 IEEE Engineering in Medicine and Biology Conference (EMBC). IEEE. (Link)
  3. Kimia Kashkoooli*, Sam L. Polk*, …, and Oluwaseun Akeju. (2019). “Drug-Specific Models Improve the Performance of an EEG-based Automated Brain-State Prediction System. 2019 IEEE Engineering in Medicine and Biology Conference (EMBC). IEEE. (Link)

* Indicates Co-First Author