Opinion
Education Opinion

CMU: Pittsburgh’s Learning Engine

By Tom Vander Ark — December 12, 2017 7 min read
  • Save to favorites
  • Print

By: Justin Aglio and Tom Vander Ark

Fifty years ago Carnegie Tech and the Mellon Institute merged to form Carnegie Mellon University (CMU). With roots in industry and enterprise, CMU grads have gone on to start tech giants Adobe, Juniper Networks and Sun Microsystems (now part of Oracle).

CMU’s early strength in artificial intelligence yielded Carnegie Learning, a leader in adaptive secondary and postsecondary math software.

CMU prof Dr. Luis Von Ahn started reCAPTCHA (now part of Google) and language learning giant Duolingo. The expertise in natural language processing resulted in more learning startups including WeSpeke and Lightside Labs (part of Turnitin).

Other education related startups include Niche (formerly CollegeProwler) and Expii, a free platform that embraces diverse learning styles to “empower you to take control of your education.” Expii’s founder, Dr. Po Shen Loh is also the national coach of the USA International Mathematical Olympiad team.

Derek Lomas, a Ph.D, graduate from CMU’s Human Computer Interaction Institute, founded Playpower Labs which helps education companies create next-generation learning products based on cognitive science, machine learning and human-centered design.

In 2002, Jesse Schell, Distinguished Professor of the Practice of Entertainment Technology at CMU, founded Schell Games, the largest full-service education and entertainment game development company in the country.

RoboTutor, a spinoff created by Professor Jack Mostow, was named one of five Global Learning XPRIZE finalists for its software that teaches children basic math and reading skills.

The CMU Center for Technology Transfer and Enterprise Creation has assisted in the formation of more than 50 companies in the last five years including high flying artificial intelligence startups.

Machine learning professor Eric Xing’s Petuum raised $93 million from SoftBank in October.

The Science of Learning

In addition to being a powerhouse in AI and tech startups, CMU is a leader in learning sciences. Dr. Ken Koedinger leads the CMU LearnLab. Koedinger created Cognitive Models, simulations of student thinking that guide the design of educational materials, practices and technologies. These models led to EdTech tools called Cognitive Tutors that support learning within rich problem-solving environments--the early basis for Carnegie Learning as well as LearnLab.

LearnLab is one of six NSF funded labs that supports basic and applied learning research. It’s also part of the CMU Simon Initiative (@SimonInitiative), named for the late Nobel and Turing prize winning professor Herbert Simon, which harnesses a cross-disciplinary approaches to measurably improve student learning outcomes.

Koedinger believes “Improvement in education will require converting teaching from a solo sport to a community based research activity.” LearnLab seeks to support and develop a new generation of learning engineers.

By studying 30 dimensions of teaching across many kinds of courses, LearnLab research led to a set of instructional principles including:

  • Personalized: Matching up the features of an instructional event with students’ personal interests, experiences, or typical patterns of language use, may lead to more robust learning compared to when instruction is not personalized.
  • Visual: Instruction that includes both visual and verbal information leads to more robust learning than instruction that includes verbal alone.
  • Guided: Instruction leads to more robust learning when it guides the learner’s attention toward relevant features.
  • Examples: Students learn more efficiently and more robustly when more frequent study of worked examples is interleaved with problem solving practice.

“Every school and university should be a LearnLab,” said Koedinger. To improve the ability for K-12 schools to make active use of learning science they will need shared data infrastructure and learning analytic methods (For more see a Global Learning Council report).

District Partnerships

CMU is known for its industry partnerships in AI, computing, and robotics. They have extended this expertise to local districts to help shape a K-12 approach to computer science. Two partners include Montour and South Fayette, two small districts west of Pittsburgh.

Almost a decade ago the districts enlisted the help of professors to lead student teams, provide after-school outreach in computer science, to engage in educational research together and to create successful strategies to support computer science K-12.

South Fayette Director of Technology and Innovation Aileen Owens (@InnovationSFSD) said, “We have become an incubator for piloting curriculum and strategies that can be implemented to all districts to provide equity in education. The CMU team has continued to connect us to other valuable partners in the computer science department, for the purpose of expanding opportunities to help us reach the next level of development in creating innovative thought leaders.”

Education Research Summit

Last month CMU and the Montour School District hosted an #EdResearchSummit. The first day was held at the Gates Center at CMU (below) and the second at Montour School District.

Dr. Ashley Coudriet, LearnLab Research Fellow (sponsored by the Grable Foundation) kicked off day one of the summit. Coudriet serves as a liaison between CMU and LearnLab school district partners (Montour School District, Cornell School District, Carlynton School District, Armstrong School District). In addition to the LearnLab districts, dozens of school district teams including superintendents, principals, and teachers participated in the summit through group discussions and roundtable presentations of best practices.

Equity was a focus of the research summit and scholarships for registration to underserved districts were provided by the Heinz Endowments.

John Balash (@J_Balash), Educational Network Coordinator at the CMU Entertainment Technology Center, kicked off the research summit introducing teams investigating augmented and virtual experiences, game-based learning through empathy and problem-solving, and digital tools promoting critical-thinking and collaboration.

Ken Koedinger presented the challenges of applying learning sciences. The first problem is that the research isn’t conclusive. Only ten percent of federally funded randomized controlled trials show significant positive results. The second problem is that experts from different fields-- cognitive neuroscience, psychology, education leadership--don’t agree.

Dr. John Stamper (@JohnCStamper), Technical Director of the Pittsburgh Science of Learning Center DataShop, spoke about forming a learning engineering community by engaging schools in research partnerships.

Stamper’s main area of research is focused on using “Big Data” from educational systems to improve learning. An expert on intelligent tutoring systems, Stamper is founder of TutorGen (@TutorGen) a scalable adaptive scoring engine. He’s also the lead researcher behind DataShop, a big repository of open data from learning systems.

Additional presentations on day one included Mixed-Virtual Reality with Dr. Nesra Yannier (@NesraYannier), AI as Cognitive Augmentation with Ken Holstein (@d19fe8), Open Learning Initiative, and Turning Research into Practice by Carnegie Learning.

Day two of the summit was kicked off by Dr. Christopher Stone, superintendent at Montour, followed by the executive director of the Grable Foundation, Gregg Behr. Tom led a discussion about the future of learning. Justin led a tour of Montour high school and elementary schools observing practitioners in action, roundtable presentations (supported by OnHand Schools) by participating attendees, and closing by Stanley Thompson of the Heinz Endowments.

The #EdResearchSummit illustrated that CMU has become one of the world’s leading engines of education innovations including machine intelligence, learning sciences, cognitive tutoring and new media. The complementary #RemakeLearning initiative has boosted regional interest in active learning and community partnerships. The two make Pittsburgh one of the most interesting education ecosystems in the country.

This post is part of a blog series in the upcoming “Getting Smart on Reinventing Education” Smart Bundle produced in partnership with The Grable Foundation. Join the conversation on Twitter using #RemakeLearning. For more, check out the other blogs in the series:

Justin Aglio is Director of Academic Achievement and District Innovation at Montour School District and Visiting LearnLab Fellow at Carnegie Mellon University. Find Justin on Twitter at @JustinAglio.


Stay in-the-know with all things EdTech and innovations in learning by signing up to receive the weekly Smart Update.

The opinions expressed in Vander Ark on Innovation are strictly those of the author(s) and do not reflect the opinions or endorsement of Editorial Projects in Education, or any of its publications.