Over 7.5 million U.S. students receive special education services under the Individuals with Disabilities Education Act, yet researchers have limited access to the Individualized Education Programs (IEPs) that define those services. IEPs are rich in detail but difficult to analyze at scale due to their unstructured, variable format and inclusion of legally protected student information. This presentation introduces two AI-based tools that address these challenges, developed under the SEAMLESS project (Special Education Applications of Machine Learning to Enhance Student Success). Zone Redactor extracts and de-identifies IEP text, while ClassifyED categorizes content—such as goals and services—by domain, grade level, and disability-related need. We report findings from a pilot partnership with a New England school serving students with specific learning disabilities, demonstrating how these tools support scalable analysis of special education data and open new possibilities for policy-relevant research.

Content is hosted by All Academic Inc, a leading provider of online hosting and software solutions for academic conferences supporting submission, peer review, scheduling, invitaions, volunteers, scheduling, web-based programs, advanced bulk email, custom workflows, and more for conferences of scholarly societies since 1999." />
BackHome
Refresh: OffView Personal Schedule
Individual Submission Summary
Share...

Direct link:

Poster #114 - Unlocking the Potential of IEP Data: Designing AI Tools for Scalable Special Education Research

Friday, November 14, 5:00 to 6:30pm, Property: Hyatt Regency Seattle, Floor: 7th Floor, Room: 710 - Regency Ballroom

Abstract

Over 7.5 million U.S. students receive special education services under the Individuals with Disabilities Education Act, yet researchers have limited access to the Individualized Education Programs (IEPs) that define those services. IEPs are rich in detail but difficult to analyze at scale due to their unstructured, variable format and inclusion of legally protected student information. This presentation introduces two AI-based tools that address these challenges, developed under the SEAMLESS project (Special Education Applications of Machine Learning to Enhance Student Success). Zone Redactor extracts and de-identifies IEP text, while ClassifyED categorizes content—such as goals and services—by domain, grade level, and disability-related need. We report findings from a pilot partnership with a New England school serving students with specific learning disabilities, demonstrating how these tools support scalable analysis of special education data and open new possibilities for policy-relevant research.

Author