Postdoctoral Researcher - Artificial Intelligence
Job no: 495773
Work type: Full-Time
Location: University City - Philadelphia, PA
Categories: Drexel University, College of Computing and Informatics
Drexel is a comprehensive global research university ranked among the top 100 in the nation. With over 24,000 students, Drexel is one of America's 15 largest private universities. Drexel is one of Philadelphia's top 10 private employers, and a major engine for economic development in the region. Drexel has committed to being the nation’s most civically engaged university, with community partnerships integrated into every aspect of service and academics.
Drexel University's College of Computing and Informatics is seeking applications for a Postdoctoral Researcher. The position is funded by the Defense Advanced Research Projects Agency (DARPA). The purpose of the funded study is to develop an Artificial Intelligence (AI) system that can automatically diagnose medical conditions from point-of-care ultrasound (POCUS) imagery. The project aims to develop novel AI approaches for leveraging domain knowledge to learn accurate models for new conditions given only a small amount of labeled training data.
The role of the Postdoctoral Researcher is to provide support in extracting POCUS domain knowledge from both medical experts and unlabeled ultrasound imagery and leveraging this knowledge to develop explainable AI models. Knowledge will be extracted from medical experts using a combination of cognitive task analysis (e.g., interviews and think aloud) and interactive machine learning (e.g., experts will interactively provide examples of key POCUS features and diagnosis explanations and feedback on the outputs of learned models). Domain knowledge will also be extracted from unlabeled POCUS imagery with unsupervised representation learning (e.g., neural network based sparse coding). Collected domain knowledge will be used within explainable AI models (e.g., simple classification models with neural-network/case-based reasoning twins to support explainability) to learn accurate classifier with just a small amount of labeled data for target medical conditions.
We are seeking candidates for a one-year appointment (renewable), with the possibility of remote work. The ideal start date for the position is 1 June 2021.
- Extract POCUS domain knowledge from both medical experts and ultrasound imagery for use in AI model development.
- Develop explainable AI models for automated diagnosis of medical conditions from POCUS imagery.
- Prepare and submit manuscripts to appropriate conferences/journals in the field.
- Prepare technical reports for internal uses as well as for submission to funding agency.
- Help manage and advise graduate and undergraduate students.
- Interact professionally with all members of the research group.
- PhD in computer science, information science, human-computer interaction, or a related area
- Strong background in artificial intelligence (particularly in machine learning methods such as representation learning and interactive machine learning, as well as knowledge representation and explainable AI).
- Experience or interest in extracting knowledge for AI models using cognitive task analysis methods (e.g., think-aloud, interviews, or interactive machine learning).
- Technical skills in statistics, writing abilities, and good communication skills.
- Typically sitting at a desk/table
This position is classified as Exempt with a salary grade of I. For more information regarding Drexel’s Professional Staff salary structure, https://drexel.edu/hr/career/ducomp/salstructure/
Special Instructions to the Applicant
Please make sure you upload your CV/resume and cover letter when submitting your application.
Additional Required Documents
- Please have two letters of recommendation sent directly to Chris MacLellan (firstname.lastname@example.org). If you would like more information or if you have any questions or inquiries, please contact Chris MacLellan.
Review of applicants will continue until a suitable candidate pool is identified.
Advertised: Eastern Daylight Time
Applications close: Eastern Daylight Time
Back to search results Apply now Refer a friend