Rebuilding a STEM Program From a Justice Lens: The Science Research Mentoring Program Meets Machine Learning (Poster 2)

• In Event: State of the Ever-Changing Field: Understanding Justice-Centered Efforts Across Informal STEM Learning Institutions

Abstract

Objectives/Purpose
The Science Research Mentoring Program (SRMP) of the American Museum of Natural History (AMNH) provides high school students from historically underrepresented backgrounds access to year-long STEM research internships alongside scientist-mentors. We are conducting a multi-year research project to teach machine learning (ML), a subset of Artificial Intelligence (AI) where computers learn patterns from data to make predictions. We designed a new ML-embedded Summer Institute, a 4-week pre-internship bootcamp introducing ML algorithms and models through natural science datasets and real-world, societal applications. We are examining how a justice framework can effect how learners see themselves through various methods that recognize their performance and competence in ML and AI contexts.

Perspective/Frameworks
ML and AI require broad participation from people who know about the social systems, phenomena, and problems that AI is meant to assist, as well as ethics and bias in AI systems. But this broadening necessitates an interrogation of whether/how ML and AI programming/learning environments support a broader diversity of learners to see themselves as central to the field of AI and as valued members of the AI community (e.g., Collins, 2018). We use an identity lens (Carlone & Johnson, 2007) to center diverse learners’ sense of self in ML.

Methods
Using a participatory design-based research approach (Getenet, 2019), we have been collaboratively working towards the vision of a justice-centered ML curriculum with Authors (Reseachers). Using concepts, frameworks, and tools from their project (blinded), we have formed a new research-practice partnership effort to apply core equitable practices (CEP) towards more just informal STEM learning experiences.

Results
First, the CEP of ''Authority Sharing'' helped deepen our understanding of the SRMPmachine Alumni Advisory Board’s role and impact in centering perspectives of young professionals pursuing computationally-demanding fields. Applying an authority sharing lens, we examine multiple levels of program design and enactment. For example, focus groups with the Board shaped the curriculum, and alumni ultimately vetted the curriculum blueprint, reviewed the final curriculum and associated assessments, and continued to engage in the program’s roll out. Second, during the curriculum design process, we reviewed the emerging SRMPmachine syllabus through CEP lenses of “Authority Sharing,” ''Shifting Narratives,” and “Critically Being With.” Through conversations and reflective dialogue, we identified pedagogical strategies aligned with these practices, and added elements to deepen impact. For example, an activity on “Who Does Science?” was reframed featuring interviews with NYC retirees-turned citizen scientists, a website created by a SRMP alumna cataloging established and emerging female science icons STEM (including SRMP alumni and mentors), and videos elevating indigenous wisdom in science. These curricular choices were a response to the question: “How are we supporting student thinking on who is a STEM thinker and doer?”

Significance: The new SRMPmachine curriculum will be tested for the first time in August of 2022, and marks one of the first ML-embedded science research curricula in the Country. We will continue our work with both the SRMPmachine Alumni Advisory Board and our collaborator team as we deliver the curriculum and iterate on its design.

Authors

• Mark Weckel
  Presenting Author

• Preeti Gupta, American Museum of Natural History
  Presenting Author

• Gabrielle Rabinowitz, American Museum of Natural History
  Presenting Author

• Kate Moore, Massachusetts Institute of Technology
  Presenting Author

• Day W. Greenberg, Indiana University
  Non-Presenting Author

• Maria Strangas, American Museum of Natural History
  Presenting Author

• Jennifer Dawn Adams, University of Calgary
  Presenting Author

• Irene A. Lee, Massachusetts Institute of Technology
  Presenting Author

• Rachel Chaffee, American Museum of Natural History
  Presenting Author