CABPortal Logo

CABPortal

The goal of CABPortal is to serve as a meeting place and repository for interdisciplinary and community-engaged collaboration and research. There are many academic collaborative models in existence today. CABPortal is home to numerous projects that utilize a collection of these models. Collaborative models being studied and implemented now are built on previous years of collaborative engagement and research.

  • Learn more about three types of distributed expertise models.
  • Learn more about how the CAB model builds on distributed expertise models.
  • See how CABPortal has impacted TCNJ students and faculty!

CAB Model

What distinguishes the CAB model from other collaborative computing models is its incorporation of a community partner into the endeavor of interdisciplinary learning. Utilizing the CAB model, two classes from different disciplines, under the guidance of their respective faculty members, work with a community partner on a community-identified problem. The CAB Model was developed over years of work that began in 2006 by educators actively researching and engaging with distributed expertise models.

To date, there have been two projects awarded NSF-grant funding that have used the official CAB Model: CABECT (Collaborating Across Boundaries to Engage Undergraduates in Computational Thinking), NSF DUE Award #1141170 and CAB (Collaborating Across Boundaries to Engage Undergraduates in STEM Learning), NSF DUE Award #1914869.

CABECT – NSF DUE Award #1141170

CABECT can be categorized as the pilot project for CAB. This project utilized the CAB model with a focus on computational thinking. Students in computer science, journalism, and interactive media worked with Habitat for Humanity to develop tools that made it easier to understand what pollutants might be on properties and associated cleanup costs. This also led to the creation of the SOAP database.

CAB – NSF DUE Award #1914869

This larger-scale project originated at TCNJ to address the growing need for STEM professionals. Its goal: increase motivation for STEM and civic engagement by using the CAB model. Classes from STEM and non-STEM disciplines paired up with community partners to work on real-world problems. Each new collaboration involved a new partner and new colleague, gradually building a network of interdisciplinary engagement.

Unlike its predecessor, CAB isn’t limited to computational thinking. It focuses on scientific literacy more broadly. Projects involved multiple faculty and community partners, and hundreds of students. Participation is open to any discipline as long as a previously-participating faculty member joins the collaboration.

Key Research Questions

Numerous collaborations are underway under the CAB research project, designed to answer questions such as:

  • Does the model scale effectively?
  • What are best practices for interdisciplinary, community-engaged pedagogy?
  • Is the model effective for STEM learning across different majors? Which ones benefit most?
  • What combinations of courses yield the strongest outcomes?
  • Does the model help historically underrepresented student groups in STEM?
  • What types of CEL projects foster the greatest STEM learning across disciplines?

All of the CAB research collaborations follow a train-the-trainer model: each new collaboration includes a returning faculty member to ensure continuity and growth.

CABAI Logo

CAB Intelligence

CAB Intelligence stands for Collaborations Across Boundaries Intelligence. It is the system that powers recommendations in CABPortal.

CAB Intelligence uses a transformer model, the same kind of AI behind tools like ChatGPT.

What does the transformer transform? It takes your words, such as interests or project keywords, and turns them into numbers called embeddings. These embeddings help us understand what your words mean in context.

We use these embeddings to find projects and people who are thinking in ways that are similar or complementary to you. This helps match you with collaborators who might be a great fit, even if you describe your work differently.

How do we do it? CAB Intelligence compares the embeddings between users and projects, and scores them based on how much they overlap and how different they are in interesting ways. These scores help rank recommendations by both similarity and novelty.

Why You See a Graph

The 3D graph you see is a visual representation of your embeddings. We show this so you can see exactly what CAB Intelligence is working with, and not just trust the AI behind the scenes.

This helps keep the system transparent. Instead of letting the AI make hidden decisions, we show you how your words were understood and compared.