Research roundup

Carnegie Mellon students and faculty are always engaged in some awesome research projects, and this section of SciTech is intended to highlight some of the work happening right on campus! Read on to learn about a few of the recent developments that might have flown under your radar.

Carnegie Mellon University students train AI to write book of limericks

With all the rhyming, witty phrasing, and creative imagery that goes into every verse, it’s hard to come up with good poetry. That’s why, as a way to assess the creativity of machines, it’s common to try and get artificial intelligence to write coherent poetry after analyzing thousands of human examples.

Using the open-source language model GPT-2, Mitch Fogelson, a Ph.D. student in mechanical engineering, along with Xinkai Chen, who completed the Master of Information Systems Management program in December 2020; Qifei Dong, a master's degree candidate in electrical and computer engineering; Christopher Dare, a master's degree candidate in information and communication technology; and Tony Qin, a junior artificial intelligence major, generated thousands of limericks after training on a database of 90,000 human-written ones. Picking the top 100 to 200 limericks from the whole collection, they have since published it in the first book of AI-generated limericks called For You, Humans. Read more about it here

Sloan Foundation provides funding for Robot Archive

Robotics has come a long way, and as with any developing area of study, it’s important to remember the historical breakthroughs and pioneers that made today's technologies possible. To better preserve the history of robotics, the Alfred P. Sloan Foundation has awarded $150,000 to the Carnegie Mellon University Libraries to fund the research into the creation of the Robot Archive, a multiphase, multiyear program to create the preeminent archive dedicated to the preservation and curation of the history of robotics. Though online for the first phase, the Robot Archive hopes to eventually hold exhibitions and conferences, as well as create a laboratory-style museum to display their collection.

This is a massive undertaking that will involve the School of Computer Science, cross-campus partners, and a team of multidisciplinary archivists and information professionals. Read more about it here.

Understanding the forces that shape an organ

The human body is a complicated biological machine replete with countless interdependent systems. Understanding these systems is crucial, especially as we look towards a future where living tissue can be bioengineered for specific use cases, such as for organ transplants.

One area that has stumped researchers is how complex three-dimensional physiological forces direct and affect organ development. Thanks to the work of biomedical engineering and materials science engineering professor Adam Feinberg, along with postdoctoral fellow Dan Shiwarski and graduate student Joshua Tashman, we now have a better understanding of such forces through their novel nanomechanical biosensor. Read more about it here.

Direct access to the human brain

The brain is one of the most important organs in the human body, being the seat of sensation, thought, reason, and emotion. As expected of an organ so incredibly complicated in its design and operation, studying it is an astronomical feat in and of itself.

That has not stopped some from trying to develop better methods for this undertaking, and recently, Maysam Chamanzar of Carnegie Mellon University and Azadeh Yazdan of the University of Washington have received a National Institutes of Health (NIH) R01 grant to create a dura smart port that will allow direct access to the brain using optical and electrical stimulation. This new technology will allow for better recording of cortical areas while minimizing the risk to exposed brain tissue. Read more about it here.