[Bytes of Good Ep. 3] Academic Research and Tech for Social Good
An edited transcript of Episode 3 of the Bytes of Good podcast. Angela Jin and James Wang talk to Dr. Aditya Vashistha about his experiences in tech for social good as an academic researcher.
Find the episode on Spotify and other streaming platforms here, and feel free to contact us at podcast@hack4impact.org if you have any questions or feedback!
ANGELA JIN: Hey there! I’m Angela Jin.
As I started my senior year of college, I had arrived at a crossroads that many of us may be familiar with: what should I do post-graduation?
On one hand, I decided that I really wanted to pursue a career in academic research.
Throughout 4 semesters as an undergrad research assistant, I had fallen in love with the process of exploring new ideas and communicating these discoveries with others through reports, papers, and presentations. Through my encounters with other researchers, I had found their seemingly insatiable curiosity to be incredibly inspiring.
At the same time, as a member of Hack4Impact, a nonprofit that builds software for other nonprofits, I’ve realized my passion for doing social good. Specifically, as a computer scientist, I’m fascinated by the rapid changes in technology and the impacts that these changes can have on society.
Considering these interests, my primary question was: How do I combine my passion for academic research in computing with my desire to do social good?
To get a better understanding of what social good looks like from the academic space, I’ve reached out to my friend, James, who has experience in building technology for social good and also in computing research. Together, we’ll speak to Dr. Vashistha, an assistant professor at Cornell University, to hear his personal perspective on the intersection between academia and social impact.
Hi James, would you mind introducing yourself?
JAMES WANG: Yeah of course! Hey Angela and listeners. My name is James, and I’m a current student at the Georgia Institute of Technology studying computer science. I’m also the current national director of Hack4Impact as well as a producer here at Bytes of Good.
ANGELA JIN: Awesome, thank you so much for being here today, James. I’m super excited to talk to you and also Dr. Vashistha, who we’ll introduce later, about academia and social impact. First, how would you describe your understanding of academia and social impact?
JAMES WANG: Yeah! I guess on the social impact side, I’ve worked with a lot of nonprofits as well as social enterprises. For the academia portion, I was a researcher at the Technology & International Development Lab at Georgia Tech. While working with nonprofits, we conducted research regarding hate speech in emerging democracies during election cycles. My specific project focused on the Myanmar election that just passed in November. I’ve also had to work with many professors and PIs of labs to learn more about how they have social impact in academia.
ANGELA JIN: Wow, that’s awesome. I’m glad we get to hear your perspectives on some of topics we’ll be discussing today. Before diving into our conversation with Dr. Vashistha, I’d like to ask you a couple background questions to ground our understanding of social impact and computing research. To begin, what even is social impact?
JAMES WANG: There are a lot of ways to define social impact.
For example, Good Finance defines social impact as the effect on people and communities that can happen as a result of action or inaction, an activity, project, program, or policy.
Each organization will have its own spin, but it really comes down to improving the community and world around you.
ANGELA JIN: I see! Improving the community and the world around you… So then, what does the phrase “technology for social impact” refer to? How does one combine technology and social impact?
JAMES WANG: For tech and social impact, they typically refer to technology being used as a tool to tackle societal challenges. In other words, tech is a medium through which you can have social impact.
When you’re building these technologies for social impact, I, and Hack4Impact, like to frame our efforts within a set of guiding principles.
A few of these principles are:
1. impact is NOT our usage metrics, meaning, the more people and the more usage does not necessarily mean we’re having great impact
2. our end users are NOT necessarily the users we want to have an effect on
3. we always want to be measuring our impact
ANGELA JIN: Got it, these guiding principles sound super helpful! I’ve heard of guiding principles for other industries such as the FinTech, EdTech, Social Media, and rideshare industries, so it’s definitely helpful to enumerate a set of guiding principles for the social impact industry. I’ll keep these ideas in mind, and hopefully our listeners also keep reflecting on these guidelines, as we continue exploring social impact in computing research.
Pivoting to research, and, more specifically, academic research, my understanding is that researchers in computing fields want to make novel discoveries that can, at some point, be used by society to improve something. Are there some research areas that are more conducive to directly achieving social impact?
JAMES WANG: Yeah, I would describe it as, how you just said: conducive. There are definitely more natural fields or labs that make working on socially impactful projects that’s easier. It’s not impossible though. Naturally, I would say, areas such as Human-Computer Interaction and Information and Communication Technologies for Development (ICTD) are easier than fields like Computer Architecture. And this is just because those areas are farther from people and society would be more difficult — but not impossible.
ANGELA JIN: It’s great to hear that you can have impact from any field, but that some may be more conducive towards having social impact than others. You mentioned HCI and ICTD as two fields that lend themselves better to conducting computing research for social good, and I believe our guest today does research at the intersection of these two areas. Could you tell me a little bit more about HCI and ICTD?
JAMES WANG: Of course, from my experience doing research in HCI, at any academic level, it’s about studying the interfaces between humans and computers, or people and computers, and better understanding how humans interact with computing technologies. HCI can be used to achieve social good through projects like creating assistive technology for people with disabilities and building educational learning for children, but those aren’t all of them. Moving onto ICTD, or Information and Communication Technologies for Development, my understanding is that this field explores how technologies can improve the lives of underserved populations.
For example, the ICTD Lab at University of Washington, or UW (“U-Dub”), does amazing work in this area, and some of their contributions include designing voice-based social media platforms for low-income, low-literate people and also producing mobile data collection tools for resource-limited settings. I definitely recommend checking out their lab page if you haven’t already, and you should definitely should learn more about all the different types of research they do!
ANGELA JIN: Wow that’s incredible! Thank you for sharing these examples of amazing research with us, and I’ll definitely be looking more into HCI and ICTD research efforts in the future. To you, our listeners, I highly encourage you to do the same if any of this sounds even remotely interesting.
Now that we’ve gained some background knowledge of social impact and its intersection with computing research, we’ll pivot to introducing our amazing guest for today’s episode.
To explore some of the ways academia can engage in social impact, we reached out to Dr. Aditya Vashistha, an assistant professor at Cornell University’s Information Science department, whose research lies at the intersection between HCI, ICTD, and Accessibility.
Before joining Cornell’s Information Science department, Dr. Vashistha completed a Ph.D. in Computer Science and Engineering at the University of Washington, where his dissertation, named “Social Computing for Social Good in Low-Resource Environments”, won the William Chan Memorial Dissertation Award.
Dr. Vashistha is also a Visiting Researcher at Microsoft Research and a faculty fellow at the Cornell Atkinson Center for Sustainability.
ANGELA JIN: Thank you so much for joining us today, Dr. Vashistha!
ADITYA VASHISTHA: Thank you for having me on the podcast. I am so excited to be here!
ANGELA JIN: Awesome. So for our first question, we heard that you started your career with industry research and we would love to hear more about your initial research experiences that led up to you becoming a professor. How did you get started with research?
ADITYA VASHISTHA: You are right Angela. After my undergrad, I worked at two different industry research labs in India before joining a Ph.D. program. My first position was as a software engineer at a multinational company, and I realized in a month after joining, that they have a research division. Frankly, I had no idea at the time how to get involved. One day, I just wrote to a senior executive on the research team, expressing that I am really enthusiastic about research and I would love to explore research opportunities. Luckily, they were looking for someone young to join the team at that time, and ideally, most of the positions in the team were for people with a Ph.D., but at that time, they were looking for someone young, and things worked out well for me, stars aligned. I was interviewed the very next day and was transferred into the research division, where I working on designing and building multimedia security systems.
While the process of learning how to do research was very fulfilling, over time, I realized that I am much more excited to use my skills to work on problems experienced by underserved communities in low-income environments, instead of designing and building solutions for the first billion people, like me, who already have too many technologies at their disposal. This desire was particularly strong because in India, the country where I was born and raised, many social ills like poverty, lack of access to education, healthcare, electricity, gender disparity, and many other things, are literally a stone-throw away.
So I started looking for research that focuses on “social good” and was extremely fortunate to find Microsoft Research in India, that had a dedicated group of researchers who were exploring topics that resonated with my research interests. I applied to work with them as an Assistant Researcher and after a series of interviews, I was hired to do something that I really wanted to do. I had a great time at Microsoft Research India. I not only got the opportunity to do research and build new computing systems aiming to make our societies more equitable and prosperous, I also learned the importance of focusing on all aspects of creating innovative solutions — from needs assessment to conceptualization to commercialization. I learned the importance of not only coming up with an interesting research idea that might work well in a lab setting, but also to deploy it into the wild, into these communities, and critically evaluate how it would scale and sustain in the long-term.
ANGELA JIN: Fascinating — I really like that you talk about scalability and sustainability in research because those are definitely top of mind for me and also the people I talk to when we discuss tech for social impact. So if you started with industry research, what led you to your current involvement in academic research?
ADITYA VASHISTHA: So, as I mentioned that I was working at Microsoft Research India, and my advisors there told me that a Ph.D. is incredibly helpful if you would like to make a career in research. And it’s not that the degree matters. What is more important is learning the craft of doing research and the journey of becoming an independent researcher — something that Ph.D. programs focus on.
And after I finished my Ph.D., I was fortunate to receive several offers and I was debating whether to go in academia or join an industry research lab. Frankly, both of these options are very lucrative. And, what really matters is what is your objective function. And at the cost of generalizing, I’d say that academia gives you more opportunity to make intellectual contributions, whereas industry gives you better opportunities of real-world impact.
I decided to join academia for a couple of reasons. First, academia gives you freedom to pursue your own research agenda. And in contrast, in industry, the sometimes the research question is short term and often focuses on bringing value to the company. My second reason is that academia gives you the opportunity to focus on all aspects of research and the eventual product development cycle. You’re involved from needs assessment to conceptualization to commercialization. And it truly feels like you’re running multiple startups at once, which is very exhilarating.
academia gives you the opportunity to focus on all aspects of research and the eventual product development cycle
Third, you get to collaborate with people who are the intellectual powerhouse — superstar faculty and smartest students. I feel like I am playing a key role in influencing thinking of our next generation leaders.
Having said that, these days, it’s not really an “industry vs. academia” question. I see several people that take advantage of the best of both worlds — best of both industry and academia. For example, in addition to being a faculty at Cornell, I am a visiting researcher at Microsoft Research India (MSR India) and have several collaborations with them on areas of mutual interest. Not only are MSR folks intellectual leaders, they are carefully exploring ways to make technology more inclusive and accessible at a larger scale and taking the needs and aspirations of underserved communities into account.
Coming back to the academic research, so far my experience has been very rewarding. It’s pretty much like running your own “research start-up”. Sometimes in the industry research settings, you are bounded to do research that brings more value to the company, whereas in academia, at Cornell, you have all the freedom to pursue “blue sky research”. For example, you define the research that you want to do, you decide what problems you will focus on, you are responsible to create your own team and bring funds to support your research and team. So, in many ways, it’s like a start-up. You focus on a specific issue, go in depths of it, do several design iterations and deployments, and when things are looking optimistic, you like to scale and sustain it. So, there are several similarities between academic research and start-up, with a difference that start-up is focusing on a product and academic research lab focuses on research.
ANGELA JIN: I’ve never compared academic research to startups, but your analogy makes a lot of sense! You mentioned the phrase “blue sky research”, but I was wondering if there are any less desirable aspects of academic research. For example, some concerns that I’ve heard surrounding academic research include needing to publish some amount of papers and also needing to receive some amount of funding. How do you view those concerns?
ADITYA VASHISTHA: This is a really important question, Angela. Personally, I see research serving two goals: the first being advancing science, and the second being advancing society.
So, let’s first talk about advancing science. A notable thing about research is its incremental nature. We all are standing on the shoulders of the giants. We all are often putting a small dent into the existing scientific knowledge and extending it a bit more than what it was before our research. And, for this reason, it is really important to share both successes and failures, to tell others what worked and what didn’t. Often, this is done through publications, but also via articles, op-eds, sharing work on different stages in the form of presentations. In other words, it is important to share our methodology, our results, and our failures, so that others can learn from it, build on it, and probably avoid the mistakes which we did.
Having highlighted the importance of having publications, many institutions are redefining what “success” really means. And, this definition of success pays less attention to the number of papers a researcher has published and focuses more on the quality of the work, the impact that work has created, and the scientific and societal boundaries that work has pushed. For example, there is no fixed formula that says, “Hey, have X number of papers and then you will graduate” or “Have Y number of publications and you will get a tenure”. And, so my advice is, do not worry about these numbers. Focus on doing impactful research that advances science and society. And, if you do so, then things will fall in place. There is no point chasing numbers — be it the number of publications or dollar amounts of funding. More often than not, this pressure of publishing X papers and getting dollars of funding is self-induced.
But since I mentioned funding, let me speak a little bit more about it too.
I see funding as a great opportunity to get an initial validation about an idea which you want to explore. To get funding, you submit a proposal that outlines your motivations, research goals, methods, expected outcomes, intellectual contribution, broader social impact of your work. And I personally really enjoy writing proposals because that’s a great way to think deeply and carefully about your research. It helps you think, helps you ask questions like why this is important? Why this is novel? How is it going to advance science and society? So, proposals are a byproduct of careful, initial investigation on a research topic. And yes, I feel great when my proposal gets funded and I feel sad when it doesn’t, but often the feedback is really useful. And that’s part of the research process — to iteratively, carefully improve something. As I said earlier, you need enough to support your research, infrastructure and students, and for different people that number looks different. So, there is no X amount of funding that is needed to get tenure or graduate from a Ph.D. program.
ANGELA JIN: Thank you for sharing your perspective on these concerns. I love your positive mindset when it comes to these aspects of academic research, and I’m sure your words on publishing to spread knowledge, and also on funding to validate your own ideas will be very useful to keep in mind for myself, and also for any other early career researchers who may also have these concerns.
Now that we’ve discussed academic research at a broad level, we’ll narrow our focus on social impact from computing research. One common thread between doing social good and conducting research is that both utilize some kind of metrics to measure impact. When it comes to measuring research impact, I’m most familiar with more technical metrics like model accuracy in machine learning, or energy efficiency and power consumption in computer systems and computer architecture research.
JAMES WANG: Right, and jumping in here to discuss metrics for social impact, measuring or quantifying social impact is an essential aspect of building technology for social good, but it’s not as simple as measuring the number of users. Going back to the guiding principles we talked about before, measuring impact is not equivalent to usage metrics like the usage metrics or the number of people on our application. One example, instead, of measuring social impact could instead be the efficiency of how we’ve improved a nonprofit’s operations. Did we improve their supply chain so that they are helping their volunteers and reaching their goals faster? Questions like that, and measuring that sort of impact is key to making sure that we are doing our job in social good.
ANGELA JIN: Awesome. So now that we’ve touched upon examples of metrics for computing research, and metrics for social impact separately, I’m curious about metrics that are used in computing research for social good. Dr. Vashistha, since your research is primarily centered around global development for underserved communities, what metrics do you usually use to gauge whether your research was successful or not?
ADITYA VASHISTHA: This is a fantastic question! So, often it comes down to how do you measure “success”, how do you measure “development”, how do you define “social justice”, how do you evaluate if it’s served or not. Of course, there are several frameworks and models that you can use to answer some of these questions. For example, you can use Development as Freedom from Amartya Sen as a framework to design, build, and evaluate your intervention or use realization-focused justice or distributive justice as a lens to look at your interventions. And these metrics could be quantitative too. For example, the number of users who use your technology, or reported knowledge gains through access to technology, or changes in behavior through your technology like medication adherence, number of follow-up visits with doctor, how many people are getting vaccinated, and so on.
… you can use Development as Freedom from Amartya Sen as a framework to design, build, and evaluate your intervention or use realization-focused justice or distributive justice as a lens to look at your interventions…
But more broadly, the answer to this question is highly context dependent — it varies from project to project. It depends on what your goals are and how you define it and how do you measure it? The success is also when millions of people are using your technology and the success is also when your paper influences the thinking of hundreds of researchers. For my work, rather than using a key set of metrics to measure success like number of users, I generally rely on a broad array of metrics that looks at things more holistically from the perspective of advancing science and society, and these metrics vary from project to project.
I generally rely on a broad array of metrics that looks at things more holistically from the perspective of advancing science and society
ANGELA JIN: So determining some kind of metric to measure social impact is key to doing social good, but the specific metrics used depend on the domain and project — there is no universal set of metrics to be applied to all situations.
JAMES WANG: Yeah, and I agree. I believe you also mentioned that there are some concrete examples that broadly apply to all computing research, like sharing publications and deploying technologies.
Actually, on the topic of deploying technologies from the academic research space, Dr. Vashistha, it’s very cool that you’re also able to have this direct impact as a researcher. However, I’m curious about the kinds of challenges you or other researchers may face when striving for impact. Specifically, in computing for global development, I’ve heard that one key challenges in designing technologies for underserved communities is to acknowledge the difference between access and inclusion. How do you perceive this challenge?
ADITYA VASHISTHA: This is a great question! I will try to illustrate the tension between access and inclusion through several examples.
Let’s take the case of smartphones.
Smartphone penetration is rapidly increasing in developing regions. Millions of low-income low-literate people can now afford to buy smartphones due to availability of cost-effective devices and data plans. So, technically they now have access to smartphones, but they are not included in the information ecology because most of the apps are not accessible to people with low literacy, most of the content online is not in their local language. As a result, they have access to a smartphone, but they are not really included.
Another example is women’s use of smartphones. Research shows that when women in the Global South own a phone, they often share their devices with their family members and use the Internet under the supervision of male family members. So they have access to phones, but do not have an agency to independently operate them. So, they have access but are not really included. In fact, most algorithms are currently designed for and designed by urban, white, affluent men and these algorithms and systems lack any representation from underserved women. So just mere access to technology is not enough. We have to think creatively and constantly about inclusion.
In fact, most algorithms are currently designed for and designed by urban, white, affluent men and these algorithms and systems lack any representation from underserved women. So just mere access to technology is not enough. We have to think creatively and constantly about inclusion.
I will now give a last example from my own research. We have designed voice-based social computing services for people who do not have smartphones and internet connectivity. So for communities who own a basic phone, a dumb phone which is only capable of making and receiving phone calls, how best can we connect such people? So we have designed voice-based services that let users call a toll-free number to record voice messages in their local language and listen to messages recorded by others. And these services find applications in diverse domains, such as health information systems, citizen journalism services, and rural information portals.
We have deployed these service in multiple countries and multiple continents, and the adoption has been great so far. Hundreds and thousands of people have used these services collectively, but when we analyzed our deployments in different countries, we noticed a pattern — that the participation of women of female users on these services were abysmal, ranging anywhere from 5% to 15% between different deployments. And when we investigated carefully why this was happening, we realized that female users on this platform experienced harassment, flirting, abusive behavior, which drove them away from the system.
So if you look at this system from the perspective of other users, for example people with limited literacy, rural communities, indigenous communities, people with vision impairment, this system is hugely successful because they are now connected to each other, they find a lot of value in the system, they have access and they are included in a platform. But if we look at the same system from the perspective of female users, technically they have access to the system — they can always call the system and be part of the social media network — but they are not really included because of the harassing behavior, because of the abusive behavior which they are experiencing. So this is another example of how access is not equal to inclusion.
JAMES WANG: Wow, I didn’t realize there was a difference, and I’m really glad that I asked that, but this example definitely clarifies the distinction between access and inclusion.
ANGELA JIN: For sure. I also wasn’t aware of this distinction between access and inclusion. I had always thought that access, or providing access to certain technologies, meant that we were automatically including people on these platforms that we’re deploying. But these examples have shown that access doesn’t necessarily ensure inclusion.
This conversation has truly been eye-opening, and thank you so much, Dr. Vashistha for sharing your perspectives and experiences in computing for global development. I’m sure this discussion about metrics and the different challenges in computing research for social good, will help both myself and other students better understand the intersection between academia and social good. For our final question, we’d love it if you could give some advice to our listeners, who may be undergrads contemplating grad school, or new grad students just beginning their research careers. What do you wish you had known before grad school?
ADITYA VASHISTHA: I wish I knew that I can be involved in research as an undergrad. I was a pretty ignorant undergrad. Although I knew that there is something called “research”, I never really delved into it. I had no idea about how to read a paper, how to pick an interesting problem, and I really wish I started doing research early on — maybe when I was a sophomore or junior. And, it’s not that there were no people around me who were doing research, but at that time, my goals were different. I was a bit risk-averse, and all I wanted was to get a software engineering position at a top company.
When I joined industry, I had the good fortune of working with amazing mentors who taught me the craft of doing research. They invested their time, energy, resources in me to help me learn ins and outs of the research process. And, I realized the importance of getting my hands dirty.
And I can’t emphasize enough how important that is and to explore things much early on. It’s not just about research, it is true for other things in life as well. Whatever you want to do, explore it, try it, don’t be worried about failing, take risks, try out different things. And don’t do just research, but intern with a start-up, a policy think-tank, an established organization, a non-profit. Find out what works for you, what drives you. Undergrad is a great time to take risks and try out different things.
Whatever you want to do, explore it, try it, don’t be worried about failing, take risks, try out different things.
For those who are aspiring to go to grad school, I would recommend that you gain some research experience before applying, and if you are fortunate enough to receive several offers, which I’m sure you will be, you should focus more on potential advisors rather than prestige of a university when making decisions. You should focus on the culture of your prospective lab, your prospective peers, and your department when making decisions between offers.
For people who are already in grad school, the biggest advice I will give is that grad school is not about taking courses and getting straights A’s. It’s more about focusing and prioritizing research and taking courses that helps you build a solid foundation and, more importantly, helps you in your research. It’s also very important to remember that grad school is not a sprint, it’s a marathon. It’s important to remember that. You can’t use the strategies for winning a 100-meter race to winning a marathon, and Ph.D. is much like a long game. It’s important to focus on quality than quantity. It’s important to go in deep rather than be too broad. It's important to have a work-life balance. It’s about setting the right priorities. Just remember that grad school is a marathon. It’s not a 100-meter race.
Just remember that grad school is a marathon. It’s not a 100-meter race.
ANGELA JIN: This podcast episode was hosted by Angela Jin and James Wang. Production by J.J. Lu. Research by Angela Jin and James Wang. Thank you to our special guest this week, Dr. Aditya Vashistha. Bytes of Good is a collaboration between Hack4Impact and Impact Labs. To learn more about the podcast or our partner organizations, visit us at bytesofgood.org. That’s bytes with a “y”. Be sure to follow the podcast @bytesofgood on Facebook, Twitter, or Instagram for show updates. Thanks again for listening.
ADITYA VASHISTHA: My biggest recommendation is that you get your hands dirty. Unless you do research, you don’t know what it really means. I often advise my undergrads that these four years are a great time to gain different kinds of experiences and figure out what is most appealing to you. I ask them to do an internship with a nonprofit organization, with a larger scale company like Google, Microsoft, Facebook, with a startup as well, to get a feel of how it’s different and fast moving, and of course, I always ask them to do research. You have four years as an undergrad, meaning four summers, which you should leverage in trying out different lines of work to figure out what is that which excites you the most.
My biggest recommendation is that you get your hands dirty.
And even with research, it’s important to try out different flavors — by working with different advisors who are doing research in different domains. For example, work on AI research for 6–9 months and then switch to HCI for another 9 months and then to Computing in Global Development. Of course, you can’t explore everything because of time constraints, but the idea is to get a wide variety of experience by engaging with different people and different labs for a longer period of time.
Finally, professors really like to interact with students. That’s the important and most fun part of this job. So, don’t hesitate to reach out to them for seeking research positions or even generally asking them questions about their research. Instead of working on research alone, just by yourself, I strongly recommend working with a faculty or Ph.D. Students. This often makes the research process more rewarding, because you have more people to discuss with, more people to brainstorm with, and get the initial handholding or mentoring that is needed.
Special thank you to Dr. Aditya Vashistha for joining us in this month’s episode!
Transcribed by https://otter.ai, funded by the Meinig Family Cornell National Scholars program.
