As stay-at-home orders swept across the country during the Covid-19 pandemic, our researchers answered the call to help others.
When the campus announced it would close, they packed up and relocated the 3D printers. They applied for grants. They adjusted ongoing research projects and they even started new ones -- all from their homes.
Here at the Human-Computer Interaction (HCI) Institute, the goal of our everyday work is to improve people’s lives through technology. HCI is a broad field, though, and improving lives can mean a variety of things as we combine our expertise in the interdisciplinary research areas of computer science, cognitive sciences and design.
We want to thank everyone at the HCII for getting through a challenging semester. Take a look below as we outline a few of our projects in response to Covid-19.
While some HCI folks were working on digital solutions to public safety, others were applying their skills to produce much needed personal protective equipment (PPE) for medical workers.
HCI Professor Scott Hudson applied his experience with 3D printing to the maker movement.
“I was, early in the crisis, pulled into a rapid response team organized out of the University of Washington which had just begun working with the Veterans Health Administration in Seattle to try to do quick production of PPE through 3D printing,” said Hudson.
“The idea was to use 3D printing which can be quickly deployed, while waiting for traditional manufacture to be brought on line. 3D printing is good in this case because it doesn't require specialized tooling (which takes time to bring on line), can be deployed in a distributed fashion (lots of people with 3D printers, each producing some), and doesn't require a new or specialized supply chain (at least for the 3D printed parts),” Hudson said.
The rapid response team ultimately produced a design for a 3D printed “Stopgap Surgical Mask.”
The idea behind their mask was to maximize the benefits of the existing supply of filter materials during the PPE shortage.
“Basically, one surgical mask can be cut into four pieces and used as an insert into this reusable 3D printed mask. This would allow quadrupling of mask supply while waiting for conventional supplies to come on line,” said Hudson.
After establishing test procedures with the VHA and FDA, and testing it in a clinical setting, their design was the first of only a few to be designated as "Reviewed for Clinical Use" by the NIH. The design is now the first item listed on the official NIH 3D Print Exchange Covid-19 Response web page.
Other contributors from CMU include Megan Hofmann, HCI Ph.D. student, and Jim McCann, assistant professor in the Robotics Institute.
Hudson contributed to several other PPE projects, too, despite the simultaneous challenges of campus closing, classes moving to online instruction mid-semester, and balancing the ongoing demands of being a CMU professor.
“I also did some work on a laser cut acrylic shield for use during intubation, a vacuformable version of the mask, and a vacuformable oxygen mask that was asked for by a NYC hospital,” Hudson said.
For a research group that works to make fabrication technology available to others, Hudson and the HCII DevLab were poised to assist with this work.
“I was already set up with four 3D printers that I use all the time here in my home workspace (plus a few experimental printers). Across my lab, when we got notice that campus was closing, we quickly dispersed much of the fabrication equipment from the lab to researcher's homes. We have always put an emphasis in our work on making things available and useful for more people, and home use is part of that. So this was not a major stretch for much of what we do,” said Hudson.
With schools closed, K-12 students across the country are relying on online learning like never before. While the pandemic is shining a light on the inequities of education in our country, it is also highlighting the benefits of face-to-face instruction for students.
Our experienced educational technology and learning sciences researchers are pivoting in response to the situation.
“Our field is on the edge of major changes. Covid-19 has forced us to rethink what instruction looks like and how (and where) students can learn,” said Nikki Lobczowski, post doctoral fellow at the HCII. “As we make these shifts, our research can help educators support students in learning mathematics.”
“There are a lot of online learning platforms to support learning math content; what makes our work especially relevant is that it's designed to support students' other needs outside of the academic content they're getting from their online courses or educational software,” added Elizabeth Richey, project scientist on PL^2 at the HCII.
“I think the current situation has highlighted just how important issues like motivation, emotion, self-regulation, and connections with teachers and peers are to students' learning, yet these kinds of issues are often left out of the digital tools available for remote learning. We hope that our work can build on the cognitive supports already being addressed through other digital tools by supporting the social, emotional, and motivational needs of students in a remote learning context,” Richey said.
“We've partnered with Carnegie Learning so that parents using the newly-available free version of MATHia@Home can monitor their child(ren)'s progress in PL^2 while also taking advantage of the resources we have developed. In addition, parents of students not working with a specific educational technology can also make use of the resources,” said Cindy Tipper, senior research programmer.
As our students’ education went online, so did many of our jobs. Researchers in our CHIMPS (Computer-Human Interaction: Mobility, Privacy, and Security) Lab immediately expanded their current cybersecurity and privacy studies in response to the unprecedented and abrupt shift to remote work.
“Working from home means, in part, having to manage your own on-site IT infrastructure,” said Cori Faklaris, HCI Ph.D. student. “They [employees] also have new ‘coworkers’ such as romantic partners, kids, pets, and roommates whose actions could impair their own cybersecurity (such borrowing secured devices and seeing confidential info, or distracting them right when they get a scam popup in their browser).”
Before the pandemic, the researchers were gathering data about how workgroups share accounts and devices, an aspect of social cybersecurity that can often be a weak link in the workplace. Now, the group is already gathering data about the transition to work-from-home, another potential vulnerability for employers and employees’ personal lives alike.
“We pivoted our interview study to add questions around how Covid-19 and the transition to home work is affecting our cybersecurity needs and pain points,” Faklaris said. The team is still seeking participants, so CMU faculty, staff and students from the Pittsburgh campus interested in participating should email them at ITstudy [at] socialcybersecurity.org to volunteer for an interview with the Social Cybersecurity Lab.
By identifying and addressing these pain points, Faklaris said the long-term goal is to develop better tools and networking for workers in non-traditional offices. They also hope to share advice for those working from home during the crisis, possibly as early as this summer.
Privacy is another HCI research area in the spotlight during the pandemic, especially in the United States.
Contact tracing (the manual process of identifying and following up with those who may have come in contact with an infected person) has long been utilized by public health professionals to control the spread of disease. However, many people are concerned about the idea of using technology, such as a public health app, to control the spread of the virus.
In order to understand how to build contact-tracing apps that more people are willing to install, the CHIMPS Lab conducted an online survey led by HCI Ph.D. student Tianshi Li.
“When reading about different countries' solutions and the heated discussion about the pros and cons of these apps in early April, I came to the realization that very few of them were actually backed up by data representing the general public's preferences, and this happens to be something I can apply my research experience in HCI and privacy to.” said Li.
User privacy is a critical design element of any system, because if users do not feel their privacy is being respected, they will not use it. And unless a lot of the population uses it, a voluntary app of this type will not be successful at tracking the virus.
“One of the main goals of our study is to figure out what design choices can strike an appealing balance between the benefits of public health and the privacy risks of user data, and lead to a higher adoption rate. This is an important problem as the effectiveness of digital contact tracing is highly related to how many people are willing to adopt them. We (CMU) are neither government nor big tech, so we can examine many possible design options without worrying about politics or the impact on the brand image. We are also naturally situated in a position that is more neutral and objective, which is important for this type of work because we do not want to bias our participants or make them think that we are biased towards any particular option,” said Li.
Initial findings show that while people in the U.S. have very diverse attitudes towards contact tracing apps, certain design choices have a significant positive effect on installation preferences, such as using centralized architecture for contact tracing and releasing infection hotspots of public areas.
“Contrary to assumptions of some prior work, we found that the majority of people (55% in our sample) are willing to install apps that use a centralized server for contact tracing, as they are more concerned about allowing tech-savvy users to identify diagnosed users, which is a fundamental risk in decentralized contract-tracing apps. There are people (25%) who seemed to be highly skeptical about contact-tracing apps in general because they did not fully trust the governments to use the data "only for good purposes" and said they ‘would never use such an app under any circumstances.’ While the rest would like to adopt at least one of the six design options presented in our survey because they found the benefits to themselves and to the society outweigh the privacy risks in some of the design options. This type of understanding could be very helpful to make informed decisions on how to build these apps,” Li said.
Learn more about their findings in their paper preprint “Decentralized is not risk-free: Understanding public perceptions of privacy-utility trade-offs in COVID-19 contact-tracing apps” or on the study’s website Towards Human-centered COVID-19 Contact Tracing Apps.
Professor Jason Hong has been interviewed several times during the pandemic to share his HCI expertise on smartphone data collection in TribLive, Covid-19 scams in TechXplore, and tech versus public good tradeoffs on NPR. He cautions us to consider both the short and long term perspectives of the situation. “There are immediate short-term needs, for example pandemic monitoring with things like contact tracing, but we also have to carefully consider what the implications will be for the long term,” said Hong.
While social distancing and stay-at-home policies are encouraged to minimize public interactions and limit the spread of Covid-19, these physically isolating measures come with social and psychological tradeoffs as well.
Bob Kraut, HCI professor emeritus, Jason Hong, HCI professor, and Tom Kamarck, professor of psychology at the University of Pittsburgh, just received a RAPID grant from the NSF to study how different types of social interactions support our social connectedness and well-being during the Covid-19 pandemic.
This timely new project builds on more than 20 years of research at the HCII and examines the influence of computer-mediated communication and online communities on well-being.
According to Kraut, "Our prior research shows that who is in the interaction, how the interaction takes place, and the activities involved may all influence whether an interaction improves well-being."
“Good mental health depends on a strong sense of social connectedness,” said Kraut. “Loneliness and social isolation are often associated with deteriorations in mental and physical health. Despite decades of research, relatively little is known about the characteristics of social interactions that lead to improvements in social connectedness and, ultimately, to mental health.”
The research collects longitudinal survey data over a three-week period to assess how participants are maintaining their social ties at this time. It will collect details about participants’ activities, including who they’re interacting with and how (eg. text, phone, video or in person).
By reaching out to participants several times per day about their interactions, they hope to learn how the quantity and characteristics of the daily interactions influence feelings of social connectedness and mental health.
In addition to advancing theory about everyday interactions and how they influence our feelings of social connectedness, the research addresses the more immediate need of understanding the consequences of social distancing policies.
For example, in situations when social interactions are limited (as they are right now), people need to know the types of activities and social interactions that are best for them. By having self-awareness about their interpersonal interactions, people can be more intentional about their activities and possibly maintain higher levels of connectedness during this and future pandemic situations.
The project will ultimately inform the development of just-in-time health-related interventions and also inform public health recommendations about the best ways to have social interactions when isolated at home.
In response to the pandemic, Carnegie Mellon unveiled five interactive COVID-19 maps that display real-time information on symptoms, doctor visits, medical tests and browser searches related to COVID-19 in the United States, including estimated disease activity at the county level.
The maps on CMU's COVIDcast website display data developed with the help of partners including Google, Facebook, Quidel Corp. and a national health system. The data, which is updated daily, will provide the general public and decision makers with a new and unique means of monitoring the ebb and flow of the disease across the country.
CMU leveraged its leadership in machine learning, statistics, data science and HCI in order to organize a variety of data and sources into 5 interactive maps in a user-friendly way.
Jodi Forlizzi, professor and director of the HCII, led the team of HCII PhD students -- Alex Cabrera, Andrew Kuznetsov, Michael Liu, and Kristin Williams -- that created the visualizations of the data sources.
“The importance of creating a good data visualization is that it makes many aspects of the Covid data and related analyses very easy for the human brain to understand. Our visualization team rapidly designed and iterated on a clear and aesthetically pleasing design that helps to convey what the data show,” said Forlizzi.
Two HCII students were involved in another local project, a contact tracing app released by CMU professor Po-Shen Lo (CMU Professor Creates Innovative App to Anonymously Trace Exposure to COVID-19).
Lo put out a call for helpers on social media and the first to answer the call was Dean Dijour (HNZ & SCS 2020).
“CMU proved to be the perfect place to find talented and passionate people with a penchant for doing impactful work on a tight timeline,” said Dijour.
“My involvement in CMU HCII, as well as broader internships and skills I've picked up over the last few years, have given me a hybrid designer/developer skill set and I was eager to put it to good use,” Dijour said.
Many of the current contact tracing apps use only Bluetooth radio waves to measure the signal strength between devices, but NOVID augments Bluetooth with ultrasound -- soundwaves with frequencies just outside the range of human hearing, which modern mobile phones can emit and record.
Dijour joined the NOVID project during his senior year at CMU, and we asked him to reflect on the daily challenges of the project.
“Every day brings unexpected challenges that require delicate, yet rapid execution,” said Dijour. “We're not afraid to jump in and spec out a radical new feature, even if it'll be scrapped. We've gone through hundreds of prototypes and variations to get to where we are now. We've had to move fast, learn from our mistakes, and push forward. NOVID is attempting to solve one of the trickiest and fastest moving problems the world has ever encountered. The problem of contact tracing is ripe with privacy concerns, sensitive medical info, even simple wording challenges. We spent sleepless nights carefully planning, testing, building, testing again; we could not stop until we had achieved a perfect technology, with simple and usable UX/UI to match.
“My days range from designing, to user testing, to coding, to working with language translators all over the world, to overseeing design implementation for cross-platform UIs. Like many people on the NOVID team, I wear many hats, and I'm very ADHD. I am constantly being pulled from one task to another, but I have to groove with it and go with the flow. This kind of mentality is what makes the NOVID team so good at moving quickly and getting things done.”
Learn more about the NOVID app at novid.org.