Landscape of opportunities at the intersection of IMU sensors and healthcare
The ubiquity of high-sampling rate motion sensor chips has presented us with the potential to transform the humble smartphone into a centerpiece of healthcare hardware democratization. Recent literature has shown initial promise in three facets of healthcare: Gait- and posture-related disorders, Geriatric care and Treatment of neurodegenerative disorders. However, these studies have been conducted in a siloed manner with small datasets and oft using low-capacity shallow machine learning models. The goal of this fellowship is to responsibly unite and standardize all of the academic-world's IMU sensor datasets into one open-sourced dataset that is mandated to specifically fuel innovations in the areas of healthcare mentioned above. In doing so, we'd like to celebrate the incredible pioneering efforts of Dr Jacquelin Perry (1918-2013).
Application deadline: Friday, October 29, 2021, 11:59 PM PST
Finalist notification: Monday, November 1, 2021
Fellowship Kick-off event: Friday, November 5, 2021, 2:00 PM PST
Fellowship Duration: November 5 - December 10, 2021. Update meetings twice a week (Hybrid attendance model)
AI Day: January 28, 2022 (Tentative)
Upon selection, the fellows will be required to participate in and finish a mandatory Kinematics-ML workshop that will be taught both on-campus (Redwood City) and streamed online which will provide a formal introduction into:
The world of motion sensor data
Human gait kinematics
A glimpse of the landscape of academic datasets
Ethics of large-scale datasets and downstream harm promulgating from cavalier dataset curation practices
After completing the workshop, the fellows will focus on:
Creating a state-of-the-art landscape of all kinematics-related motion sensor datasets.
Authoring a common pre-processing pipeline to unite all the datasets into a single monolithic meta-dataset with the right format and documentation.
Responsibly disseminating this in a peer-reviewed academic venue alongside a datasheet.
We welcome applications from practitioners, tech-enthusiasts, auto-didacts, students (at the undergraduate and graduate levels), and medical health professionals alike! We do expect some demonstrable fluency in being able to code in Python, an initiation to the ETL workflow, and some knowledge of one or more of the following basic frameworks/libraries: Scikit-learn, Pandas, SciPy, and RAPIDS-AI.
As proud signatories of the Tech Inclusion Pledge, we are firmly committed to creating and nurturing a culture of inclusivity. Hence, we would like to invite applications from individuals with diverse backgrounds and experiences to apply to our fellowship program. Given the under-representation of women and minorities in the domains of Machine Learning and Data Science, we would like to especially encourage individuals from those groups to apply.