AI-ALOE Members Receive $25K Grant at VanderbiltNov 1, 2022 — Atlanta, GA
Scott Crossley, a professor of Special Education at Vanderbilt University and Langdon Holmes, a graduate student in the Department of Psychology and Human Development at Vanderbilt University received a $25,000 grant to build out a personally identifiable information (PII) dataset.
The grant provided by the Learning Agency, a lab that aims to improve education delivery systems to ensure all learners have ample opportunities to succeed. Funds are expected to cover:
- De-identify 14,000 writing samples and release a shared dataset for educational data mining. We will label and obfuscate identifying information from 33,000 student-generated writing samples. All PII will be replaced with artificially generated replacements (original identifiers will be replaced with invented text).
- Develop models and methods to automatically deidentify student-generated text. We will develop methods for the automatic deidentification of student-generated text. We will share our source code publicly, in addition to reporting our findings.
- Raise interest in the task of automatic deidentification. Before being fully released to the public, a portion of the dataset will be shared with data science teams who will compete to develop automatic deidentification systems. These systems will be tested against the held-out portion of the data, and prize money will be awarded to the best performing systems. Source code for all competition submissions will be public.
“We are excited about expanding this work,” Crossley wrote.
