A Cornell-led research team has discovered that the algorithm behind Google Ads charged significantly more to deliver online ads to Spanish-speaking people about the benefits of SNAP, formerly known as food stamps.

Together with a survey of 1,500 Americans that found broad support for a more equitable approach to promoting SNAP, the findings led to changes to directly target more Spanish-speakers in California who seek help paying for food.

“SNAP is a really important resource to get right,” said Allison Koenecke, assistant professor of information science in the Cornell Ann S. Bowers College of Computing and Information Science. “When faced with an algorithm that has disparate impact, our research asks, how do you pick a strategy to interact with the algorithm to equitably recruit SNAP applicants?”

Koenecke is the lead author of “Popular Support for Balancing Equity and Efficiency in Resource Allocation: A Case Study in Online Advertising to Increase Welfare Program Awareness,” which was presented at the AAAI Conference on Web and Social Media in June.

“Our goal was to go beyond just quantifying these ad cost disparities,” she said. “We also wanted to address questions asked by real-life decision-makers, like what do we actually do about this? How do people actually want to allocate these ads?"

Californians can apply for SNAP benefits using a website called GetCalFresh, which is developed and managed by Code for America, a civic tech nonprofit that builds digital tools and services for community leaders and governments. Code for America primarily recruits GetCalFresh applicants through Google Ads – for example, spending roughly $400 daily to reach anyone from San Diego County who punches key words and phrases like “how to apply for food stamps” into Google.

However, despite GetCalFresh being offered in multiple languages, Spanish-speakers were filling out proportionally fewer applications than English-speakers. In San Diego County, 23% of families living below the poverty line speak Spanish as their primary language, and yet just 7% had applied for SNAP via GetCalFresh, researchers said.

Koenecke and her collaborators discovered one possible reason: the default, dollar-stretching algorithm behind Google Ads was working too efficiently and disregarding Spanish-speaking people in the process.

When Google Ads is configured to garner the most SNAP enrollments per dollar, it ends up delivering fewer ads to prospective Spanish-speaking applicants because such ads cost more than those for English speakers, the team found. At the time, for every $1 spent on Google Ads to “convert” an English-speaking applicant into a SNAP benefits holder, it cost $3.80 to convert a Spanish-speaking person – nearly four times more. Another bidding option on the Google Ads platform cost 1.4 times more to reach Spanish-speakers versus English-speakers.

Koenecke and her collaborators can’t definitively explain the difference, since Google Ads is a black box – a proprietary machine-learning tool outside of public review. It could be attributed to any number of factors, like supply and demand or a bug in the system, she said.

For GetCalFresh, the research findings pose an important ethical question regarding how to spend its limited online advertising budget: Should they reach as many Californians as cheaply as possible, even if that means fewer Spanish-speaking applicants, or advertise more to Spanish-speakers, even if that yields fewer total applicants?

Trade-offs such as these are at the heart of Koenecke’s research into fairness and algorithmic systems, which are increasingly being used to help with decision-making in areas with real consequences, like health care, banking and child services. But without additional scrutiny, algorithms – including a seemingly harmless one behind an advertising platform – can exacerbate inequality or produce results that run counter to what people actually want or need, she said.

Koenecke and her collaborators asked the public to weigh in, surveying roughly 1,500 Americans on how they would balance efficiency and equity in advertising SNAP benefits. Across age groups, gender, race, welfare status and even political party affiliation, respondents generally preferred reducing total enrollments to facilitate more enrollments among Spanish speakers, researchers found. When they ran the survey among Code for America staff, they found similar results – a strong desire for equitable access.

“If you compare the fact that the majority of both Republicans and Democrats actually prefer some amount of equity in this particular case, then this opens the door to more bipartisanship in thinking about how fairness can play a role in online applications involving the algorithmic distribution of goods,” Koenecke said.

As a result of the team’s findings, Code for America adjusted its online advertising strategy to directly target more Spanish-speaking prospective applicants.

“It’s important for the field and the public to have productive dialogues about the kinds of metrics we should be using in these algorithmic systems,” she said. “The communities most impacted by the algorithms should be given more power in the decision-making process.”

Along with Koenecke, the paper’s co-authors are Eric Giannella of Code for America, Robb Willer of Stanford University, and Sharad Goel of Harvard University’s Kennedy School.

This research was partly funded by the National Science Foundation and Stanford University.

By Louis DiPietro, a writer for the Cornell Ann S. Bowers College of Computing and Information Science.

Sarah Riley, a doctoral student in the field of information science, has received a 2023 Horowitz Foundation for Social Policy Award for her research into algorithmic pretrial risk assessments.

Riley, who studies municipal algorithmic systems, race, racism, and inequality, is one of 25 scholars whose 2022 application was selected. The Foundation received more than 550 applications.

Her project, “Pretrial Risk Assessments as Organizational Processes” focuses on the administration of pretrial risk assessments in Virginia. These risk assessments are algorithmic tools that make recommendations as to whether a defendant should be released or detained until trial. She uses a mixed-methods approach to understand how human discretion in the pretrial process –particularly on the part of pretrial officers –affects risk scores, pretrial detention decisions, and life outcomes for accused people.

Riley is expected to receive her doctoral degree from Cornell in August. This fall, she will attend Stanford University as an IDEAL Provostial Fellow. The three-year program is designed to support the work of early-career researchers who will lead the next generation of scholarship in race and ethnicity, according to Stanford.

Riley holds a bachelor’s degree in political science from Amherst College and a master’s of public policy from the University of California, Berkeley. She is advised by Karen Levy, associate professor of information science and associate member of the faculty of Cornell Law School, Solon Barocas, adjunct assistant professor of information science, and Martin Wells, the Charles A. Alexander Professor of Statistical Sciences with additional appointments in the ILR School, the College of Agriculture and Life Sciences, Cornell Law School and Weill Cornell Medicine.

Nine innovative faculty members and doctoral students from the Cornell Ann S. Bowers College of Computing and Information Science were selected to receive the second round of grants from the five-year, multimillion-dollar strategic partnership between the college and LinkedIn.

This year’s awards will help fund high-impact research projects on topics ranging from large language models (LLMs) and recommender systems to dynamic information retrieval and algorithmic fairness.

Now in its second year, the Cornell Bowers CIS-LinkedIn partnership – a first for both entities – continues to make good on its primary goals: provide annual funding to Cornell Bowers CIS to support doctoral students; engage faculty through research support; amplify efforts in diversity, equity, and inclusion; and establish a research connection between scientists and engineers at LinkedIn and faculty and students in Cornell Bowers CIS, one of the top artificial intelligence programs in the U.S.

Faculty award winners:

●      Kilian Weinberger, professor of computer science, and Yoav Artzi, associate professor of computer science at Cornell Tech. In their project, “Real-Time Differential Search Indices,” the duo will explore Differential Search Indices (DSI) for constantly changing dynamic content – an open problem in information retrieval. DSI are considered a paradigm shift in information retrieval but are currently limited to retrieving static content only.

●      Sarah Dean, assistant professor of computer science. User-provided feedback, often in the form of clicks, has powered the last decade of search and recommendation algorithms and provides platforms key insights on user engagement. However, a focus on short-term engagement metrics can lead to long-term issues like unfairness or poor platform health. In “Reliable and Efficient Recommendation & Search with Long-Term Objectives,” Dean intends to develop algorithms that make decisions on the short term while ensuring long-term platform health.

●      Mor Naaman, professor of information science at Cornell Tech. Naaman’s project, “Writing with AI, Responsibly: Understanding the Potential of LLMs to Shift Our Writing, and How to Mitigate It,” will build on research into responsible AI by providing a deeper understanding of how autocomplete functions – powered by new LLMs – can transform how we express ourselves and even change our opinions.

●      Cristian Danescu-Niculescu-Mizil, associate professor of information science. Humans have an intuition of when our conversations are heading south. In “Assessing the Health of Ongoing Dialogs,” Danescu-Niculescu-Mizil will develop methods for endowing artificial systems with this kind of intuition and will explore how this artificial intuition can be used to improve the way we communicate with each other, as well as with dialog systems.

Awards to doctoral students include academic year funding and discretionary funds. Awardees are:

●      Princewill Okoroafor, a doctoral student in the field of computer science and advised by Robert Kleinberg, studies theoretical aspects of machine learning, especially in learning theory and reinforcement learning. His project aims to advance our understanding of limitations and trade-offs in predictive models as well as develop new algorithms that can achieve more accurate and calibrated predictions – all of which can have a significant impact on the job search industry.

●      Jonathan D. Chang, a doctoral student in the field of computer science advised by Wen Sun, explores imitation learning, reinforcement learning, representation learning, and their intersection with generative models. His project, “Chatbot Recommendation Systems,” will investigate algorithms that can help reduce the amount of data needed to train reinforcement learning systems and improve responses in chatbot recommendation systems.

●      Benjamin Laufer, a doctoral student in the field of information science advised by Helen Nissenbaum and Jon Kleinberg, is interested in the values and politics embedded in technology systems used to make high-impact decisions. In “Efficient, Equitable Choices in Dynamic Environments,” Laufer will develop methods for identifying, measuring, and mitigating the harmful dynamics that arise in algorithmic recommendations.

●      Shira Mingelgrin, a doctoral student in the field of statistics and data science advised by Sam Wang, studies causal modeling. Mingelgrin’s project will apply causal discovery methods to investigate potential differences in the causal structure under different settings, for instance, testing whether the factors that drive hiring behavior vary among genders.

By Louis DiPietro, a writer for the Cornell Ann S. Bowers College of Computing and Information Science.

Researchers from Cornell Tech and the Cornell Ann S. Bowers College of Computing and Information Science are part of the first cohort of participants from four institutions to receive funding from the Google Cyber NYC Institutional Research Program, a Google-funded initiative to improve standards of online privacy, safety and security, and to establish New York City as the epicenter of cybersecurity research.

In partnership with Google, seven projects from Cornell faculty have been selected that combine basic research with computer innovations to expand our understanding of – and provide solutions to – cybersecurity issues in society. Each team will work with a Google sponsor and receive funding and access to Google Cloud Platform credits for up to three years.

“Through support from Google and in collaboration with our partners, New York will emerge as the hub for cutting-edge exploration and innovation in the fields of cybersecurity, trust, and safety,” said Greg Morrisett, the Jack and Rilla Neafsey Dean and Vice Provost of Cornell Tech and principal investigator for Cornell.

Along with Cornell, three other schools will receive funding through the program: City University of New York, Columbia University’s Fu Foundation School of Engineering and Applied Science and New York University’s Tandon School of Engineering. The Google Cyber NYC Institutional Research Program is part of the $10 billion cybersecurity initiative announced in 2021.

“Google is dedicated to forward-thinking and responsible AI development,” said Phil Venables, Google Cloud chief information security officer. “We are excited to partner with these leading institutions in AI through the Google Cyber NYC Institutional Research Program to address the ever-evolving threat landscape in cybersecurity.”

Beyond advancing digital technologies that increase digital trust and safety, the funding will also enable Cornell Bowers CIS to expand its leadership in the field of cybersecurity and increase the number and diversity of qualified cybersecurity professionals entering the workforce.

The inaugural projects from Cornell are:

• Fred Schneider, Samuel B. Eckert Professor of Computer Science: “Towards an Applied Theory of Information Flows.” Traditionally in cybersecurity, researchers have devised ways to suppress information flows, such as data leaks. This project will develop a theory for surfacing information flows in systems so they can be used to establish accountability and to determine effective means to control their content.

•  Noah Stephens-Davidowitz, assistant professor of computer science: “Foundations of post-quantum cryptography.” This work focuses on the theoretical foundations of post-quantum cryptography – cryptographic methods that enable a traditional computer to fend off attacks from a quantum computer (one that uses the quantum states of subatomic particles to store information and that can solve especially complex problems). This work is becoming increasingly important as quantum computing technology advances.

• Andrew Myers, professor of computer science: “Making Hardware Comprehensively Secure Against Spectre – by Construction.” Recently, there have been several prominent timing attacks aimed at hardware in which attackers steal sensitive information, such as passwords, by analyzing the time required to perform computations. This project will develop a new secure hardware description language to build processors that are immune to this class of attacks.

• Nicola Dell, associate professor of information science at the Jacobs Technion-Cornell Institute at Cornell Tech, and co-PI Thomas Ristenpart, professor of computer science at Cornell Tech: “Improving Account Security for At-Risk Users.” Many online services employ user-facing account security interfaces (ASIs) to show security-related information, such as recent logins and connected devices. But ASIs can be compromised, putting users – especially survivors of intimate partner violence, journalists and refugees – at risk. The researchers propose to develop more secure ASIs that preserve privacy and offer better user experience.

• Dell and Ristenpart, in collaboration with researchers at NYU: “Discovering and Locating Unwanted IoT Devices for Survivors of Intimate Partner Violence.” Abusers use Internet of Things (IoT) devices, like location trackers and hidden cameras, to surveil and harass their victims. This research aims to build tools that can discover and locate IoT devices, create a guide to train survivors on how to use the tools, and test protocols for deploying them with survivors in NYC.

• Nate Foster, professor of computer science, in collaboration with researchers at NYU: “ProSecco: Programmable In-Network Security.” As the reach of networks has expanded from servers and desktops to phones, cars, IoT devices, and cameras, attacks perpetrated through networks have increased in scale, sophistication, and frequency. The researchers propose ProSecco, a project that will use programmable network devices to create more proactive, dynamic, and automated defenses for network security.

• Kevin Ellis, assistant professor of computer science, with co-PIs Owolabi Legunsen, assistant professor of computer science, and Alexandra Silva, professor of computer science: “Safe Program Generation and Deployment by Large Language Models.” AI systems have tremendous potential for managing a range of problems and tasks in everyday life – such as dispensing medication in a hospital or handling employee calendars – but there is still the possibility of dangerous mistakes. In this project, researchers propose a key safety property, and propose developing new techniques for eliciting, specifying, validating, and monitoring this safety property for machine learning systems.

Patricia Waldron is a writer for the Cornell Ann S. Bowers College of Computing and Information Science.

Steven Jackson believes a university should be, first and foremost, about educating students, and that core belief will inspire him in his new role as vice provost for academic innovation (VPAI).

“My interest in the position really came from my central commitment as a teacher at Cornell,” said Jackson, professor in the Department of Information Science in the Cornell Ann S. Bowers College of Computing and Information Science, with a dual appointment in the Department of Science and Technology Studies in the College of Arts and Sciences (A&S).

“Teaching is at the center of what we do,” Jackson said. “For me, teaching and research go together; they are very tightly intertwined.”

Jackson’s term officially begins July 1, although he’s been conducting meetings and easing into the role for several weeks, he said.

“Steve is passionate about his teaching and helping students realize their full potential through his work, both in the classroom and in the lab,” Provost Michael I. Kotlikoff said. “He is also a conscientious university citizen whose commitment to continually improving Cornell’s teaching and learning environment for students and faculty makes him an ideal fit for this role.”

The position, created in 2017 along with the Center for Teaching Innovation, was initially held by Julia Thom-Levy, professor of physics (A&S) and associate vice provost for physical sciences. John Siliciano ’75, professor of law in Cornell Law School and former deputy provost, has been in the role on an interim basis since October 2022.

“I also want to thank Julia for her dedication in five years as our inaugural VPAI, as well as John for stepping in on an interim basis,” Kotlikoff said. “So much has been accomplished since the center was launched and Julia’s leadership and vision have helped keep Cornell at the forefront of pedagogical innovation.”

While some people might equate innovation with technology, Jackson said, that’s only part of the story.

“Technology is a tool, to be sure, but for me, innovation in teaching is really about inviting and empowering a group of committed teachers to use their imagination, their creativity and the tools around them to think and do teaching in a different way,” said Jackson, whose fields of study include technology ethics, law and policy, and human-computer interaction, including the study of how users and groups collaborate around new computational tools and infrastructures. “And we should invite students to also engage in that process and be part of the ways in which we reinvent and reimagine teaching at Cornell.”

Jackson, formerly the chair of information science – as well as house professor-dean of Keeton House on West Campus from 2015-21 – hopes to engage faculty and students and do “a lot of listening” as he grows into his new role.

“The diversity and range of teaching that goes on at Cornell has become increasingly apparent to me. There is not a one-size-fits-all approach,” he said. “There are deep disciplinary traditions of teaching, and exciting examples of innovation and experiment that have been going on for decades.

“There’s also no one-size-fits-all when it comes to students, who aren’t just blank slates or empty vessels waiting to be filled,” he said. “Our students are coming from all kinds of backgrounds and perspectives, and it’s incumbent on us to bring this in as a resource and opportunity for learning.”

Jackson received his bachelor’s in English and creative writing in 1994 from Concordia University in Montreal; his master’s in political economy in 1999 from Carleton University in Ottawa; and his Ph.D. in communication and science studies in 2005 from the University of California, San Diego.

He spent six years on the faculty at the University of Michigan before joining Cornell in 2011.

By Tom Fleischman, Cornell Chronicle

This story was originally published in the Cornell Chronicle.

Lawrence Blume, the Distinguished Arts and Sciences Professor of Economics and professor of information science, has been named the inaugural associate dean for academic affairs for the Cornell Ann S. Bowers College of Computing and Information Science beginning July 1, 2023.

During the three-year term, Larry will oversee academic affairs in the college related to faculty and academic staff. Primarily, this will include managing tenure and promotion cases, reviews and reappointments, and other academic affairs.

“Larry is an outstanding member of the faculty and brings significant research, teaching, and administrative leadership experience to this new and important role for the college,” said Kavita Bala, dean of Cornell Bowers CIS. “Having worked with him closely this past year, I have full confidence he will contribute immensely in this new position.”

Blume received a B.A. in economics from Washington University and a Ph.D. in economics from the University of California, Berkeley. He was one of the general editors of The New Palgrave Dictionary of Economics, 2nd edition, to which he contributed several articles on economic theory. Blume is also a fellow of the Econometric Society, a visiting research professor at the Institute for Advanced Studies in Vienna (IHS), and has been a member of the external faculty at the Santa Fe Institute, where he served as co-director of the Economics Program and on the institute's steering committee.  He also served as chair of Cornell’s Department of Economics. 

“It is an exciting time in the college, and I’m honored to accept this role and to work closely with faculty,” said Blume. “This role presents an incredible opportunity to shape the academic landscape and ensure Cornell Bowers CIS remains a leading academic institution.”

In addition to his administrative duties as associate dean for academic affairs, he will continue in his leadership role as the Charles F. and Barbara D. Weiss Director of Undergraduate Studies for the Department of Information Science. He will also continue to teach and conduct research in general equilibrium theory and game theory, as well as in income and wealth distribution and network design.

Blume has been serving as interim associate dean for education for the 2022-2023 academic year while Claire Cardie, the Joseph C. Ford Professor of Engineering in the departments of computer science and information science, was on sabbatical.

A research team co-led by Cornell found that for schools without the resources to conduct learning analytics to help students succeed, modeling based on data from other institutions can work as well as local modeling, without sacrificing fairness.

“To use data-driven models, you need data,” said Rene Kizilcec, assistant professor of information science in the Cornell Ann S. Bowers College of Computing and Information Science. “And in many schools, especially lower-resourced schools that would benefit the most from learning analytics applications, data is rarely accessible.”

Kizilcec is a senior author of “Cross-Institutional Transfer Learning for Educational Models: Implications for Model Performance, Fairness, and Equity,” to be presented at the Association for Computing Machinery Conference on Fairness, Accessibility and Transparency (ACM FAccT), June 12-15 in Chicago. The lead author is Josh Gardner, doctoral student in computer science at the University of Washington.

Kizilcec and his team used anonymized data from four U.S. universities, and converted it into a common structure for the purpose of modeling which students are likely to drop out of college. Only the university-specific models – no individual student data, which raises privacy issues -- were shared between members of the research team.

More than 1 million students drop out of college each year in the U.S.; they are 100 times more likely to default on their student loan payments than those who graduate. This has led the federal government to impose regulations that incentivize colleges and universities to reduce dropouts by requiring them to report dropout rates, as well as rankings that account for graduation rates.

Kizilcec said that major institutions have the resources to conduct predictive data analytics. But institutions that could most benefit from that data – smaller colleges or two-year institutions – typically don’t.

“They have to rely on the services of a few companies that offer education analytics products.” he said. “Institutions can either build their own models – a very expensive process – or purchase an analytics ‘solution,’ with modeling that is typically done externally on other institutions’ data. The question is whether these external models can perform as well as local models, and whether they introduce biases.”

The goal of the researchers’ work was an accurate prediction of “retention” – whether each student who enters an institution for the first time in the fall would enroll at that same institution the following fall.

To assess the success of transfer learning – taking information from one institution and using it to predict outcomes at another – the team employed three approaches:

• Direct transfer – a model from one institution is used at another;
• Voting transfer – a form of averaging to combine the results of several models (“voters”) trained at disparate institutions to predict outcomes at another; and
• Stacked transfer – combining the predictions of models trained on all available institutions with the training data of the source institution.

The researchers used the three transfer methods, along with local modeling at each of the four institutions, in order to assess the validity of transfer learning. Predictably, local modeling did a better job of predicting dropout rates, “but not by as much as we would have thought, frankly, given how different the four institutions are in size, graduation rates and student demographics,” Kizilcec said.

And in terms of fairness – the ability to achieve equivalent predictive performance across sex and racial subgroups – the modeling performed well without sacrificing fairness.

Kizilcec said his team’s results point to more equity in dropout prediction, which could help lower-resourced schools with earlier intervention and preventing student departures, which cost the institution and can lead to worse outcomes for the students.

“It may not be necessary after all to allocate resources to create local models at every single school,” he said. “We can use insights from schools that have data infrastructure and expertise to offer valuable analytics to schools without these resources, and without sacrificing fairness. That’s a promising result for school leaders and policymakers.”

Other contributors are Christopher Brooks, assistant professor at the University of Michigan School of Information; Renzhe Yu, assistant professor of learning analytics and educational data mining at Columbia University; and Quan Nguyen, instructor of data science at the University of British Columbia.

Support for this work came from Google and Microsoft.

By Tom Fleischman, Cornell Chronicle

This story was originally published in the Cornell Chronicle.

Friends, family, faculty, and staff filled Barton Hall to cheer on members of the newest class to join the ranks of Bowers CIS alumni, as the students traversed the stage at three recognition ceremonies held May 26, by the Cornell Ann S. Bowers College of Computing and Information Science.

More than 1,100 undergraduate, master's, and Ph.D. students received degrees this week, after completing their education in August 2022, December 2022, or May 2023. This year, undergraduates from Cornell Bowers CIS made up about 20% of Cornell's 2023 graduating class. 

In her opening remarks at the information science and computer science ceremonies, Kavita Bala, dean of Cornell Bowers CIS, recounted the remarkable history of the college. Bala recalled that 24 years ago, Cornell was a pioneer in creating the faculty of computing and information science, "envisioning multiple fields of the information age together under one banner: computer science, information science, and statistics and data science." She urged the new alumni to take that innovative history with them as they embark on their careers.

"Regardless of where your path takes you — industry, graduate studies, academia, or entrepreneurship — I encourage you to take a page from our college’s history by taking risks and looking out for any opportunity to explore uncharted waters," she said.

Statistics and Data Science

In his remarks at the Department of Statistics and Data Science ceremony, Martin Wells, the Charles A. Alexander Professor of Statistical Sciences, encouraged the more than 100 graduates in attendance to take what they have learned at Cornell and apply it to have a positive impact on the world.

"As ambassadors of statistical science, your skills will influence how our data-driven society responds to current and future challenges," said Wells. "As you step into the next chapter of your lives, remember to embrace new challenges, continue learning, and pursue your passions…May you revel in your search for clarity, understanding, and truth."

Information Science

“The world really needs you,” said David Williamson, chair of information science and professor in the School of Operations Research and Information Engineering, in his remarks. “We need people who can think about the effects of technologies before they get built, rather than after the fact. We need people who think critically about technology to know how it actually works, so that the critique can have a real impact. … I think your choice to major in information science will serve you and all of us very well in the years to come.”

Jeff Rzeszotarski, assistant professor of information science, was the faculty speaker selected by the graduating class. With a penchant for big numbers, Rzeszotarski encouraged the roughly 300 information science graduates to consider the estimated 10,000 hours they each put in to earn their diploma – about 2 million hours total.

“I call out these numbers because it's often tempting for students to focus only on the final grades that show up on a transcript,” said Rzeszotarski. “While today, we celebrate your graduation and earning your diploma, don't lose sight of all the sheer effort you had to put forth to get to this transition point.” 

He characterized information science as a field that examines how technology merges with different aspects of our world – from communication and law to art and social change – to form an interconnected, complex web that shapes society. He urged students to find connections or areas of inquiry in the world to apply their Cornell education.

“Identify where your knowledge transfers in new fields. See unexpected relationships between domains. Find these connection points in the next phase of your life,” he said. “These connections will help you solve problems in ways no one else can.” 

Computer Science

In the last and largest event of the day, the Department of Computer Science recognized more than 700 undergraduate and graduate students, many who were cheered on by supporters with banners, face cutouts, and noisemakers.

Éva Tardos, the Jacob Gould Schurman Professor of Computer Science and department chair, recognized the difficulty of attending college during the COVID-19 pandemic, and congratulated the students on overcoming these challenges.

"I hope you will go after the many great opportunities that our professional industry offers you," said Tardos. "This is a truly exciting time to be a computer scientist." She also urged the graduates to be thoughtful in their career pursuits, taking into account issues of privacy, accountability, and fairness. 

In closing, Tardos encouraged the new graduates to come back and visit and to stay in touch with the faculty. "I hope to personally hear from every one of you what your next adventure is," she said. "All 700 of you!" 

By Lou DiPietro and Patricia Waldron, both writers for the Cornell Ann S. Bowers College of Computing and Information Science.

An experiment in which two people play a modified version of Tetris – the 40-year-old block-stacking video game – revealed that players who get fewer turns perceive the other player as less likable, regardless of whether a person or an algorithm allocates the turns.

“We expected that people working in a team would care if they are treated unfairly by another human or an AI,” said Malte Jung, associate professor of information science in the Cornell Ann S. Bowers College of Computing and Information Science, whose group conducted the study.

Most studies on algorithmic fairness focus on the algorithm or the decision itself, but Jung sought to explore the relationships among the people affected by the decisions.

“We are starting to see a lot of situations in which AI makes decisions on how resources should be distributed among people,” Jung said. “We want to understand how that influences the way people perceive one another and behave towards each other. We see more and more evidence that machines mess with the way we interact with each other.”

Houston B. Claure, M.S. ’20, Ph.D. ’23, is first author of “The Social Consequences of Machine Allocation Behavior: Fairness, Interpersonal Perceptions and Performance,” published April 27 in Computers in Human Behavior. Claure earned his master’s and doctorate in mechanical engineering, minoring in computer science.

Jung and Claure conducted an earlier study in which a robot chose which person to give a block to, and studied the reactions of each individual to the machine’s allocation decisions.

“We noticed that every time the robot seemed to prefer one person, the other one got upset,” said Jung, director of the Robots in Groups Lab. “We wanted to study this further, because we thought that, as machines making decisions becomes more a part of the world – whether it be a robot or an algorithm – how does that make a person feel?”

Because of the time it took to conduct experiments and analyze data using a physical robot, Jung and Claure felt there was a better and more efficient way to study this effect. That’s when Tetris – originally released in 1984, and long a useful tool for researchers looking to gain fundamental insights about human cognition, social behavior and memory – entered the picture.

“When it comes to allocating resources,” Claure said, “it turns out Tetris isn’t just a game – it’s a powerful tool for gaining insights into the complex relationship between resource allocation, performance and social dynamics.”

Using open-source software, Claure – now a postdoctoral researcher at Yale University – developed a two-player version of Tetris, in which players manipulate falling geometric blocks in order to stack them without leaving gaps before the blocks pile to the top of the screen. Claure’s version, Co-Tetris, allows two people (one at a time) to work together to complete each round.

An “allocator” – either human or AI, which was conveyed to the players – determines which player takes each turn. Jung and Claure devised their experiment so that players would have either 90% of the turns (the “more” condition), 10% (“less”) or 50% (“equal”).

The researchers found, predictably, that those who received fewer turns were acutely aware that their partner got significantly more. But they were surprised to find that feelings about it were largely the same regardless of whether a human or an AI was doing the allocating.

One particularly interesting finding: When the allocation was done by an AI , the player receiving more turns saw their partner as less dominant, but when the allocation was done by a human, perceptions of dominance weren’t affected.

The effect of these decisions is what the researchers have termed “machine allocation behavior” – similar to the established phenomenon of “resource allocation behavior,” the observable behavior people exhibit based on allocation decisions. Jung said machine allocation behavior is “the concept that there is this unique behavior that results from a machine making a decision about how something gets allocated.”

The researchers also found that fairness didn’t automatically lead to better game play and performance. In fact, equal allocation of turns led, on average, to a worse score than unequal allocation.

“If a strong player receives most of the blocks,” Claure said, “the team is going to do better. And if one person gets 90%, eventually they’ll get better at it than if two average players split the blocks.”

Rene Kizilcec, assistant professor of information science (Cornell Bowers CIS) and a co-author of the study, hopes this work leads to more research on the effects of AI decisions on people – particularly in scenarios where AI systems make continuous decisions, and not just one-off choices.

“AI tools such as ChatGPT are increasingly embedded in our everyday lives, where people develop relationships with these tools over time,” Kizilcec said, “how, for instance, teachers, students, and parents think about the competence and fairness of an AI tutor based on their interactions over weeks and months matters a great deal.”

The other co-author is Seyun Kim ’19, M.S. ’21, a doctoral student in human-computer interaction at Carnegie Mellon University.

The work was supported by the National Science Foundation.

By Tom Fleischman, Cornell Chronicle

This story was originally published in the Cornell Chronicle.

For Zachary Schecter ‘23, it was winning the Most Innovative Game award at the Game Design Showcase his sophomore year that really clinched it: he would become a professional game developer. He watched showcase attendees play his team’s game, Sisyphus, where the mythological main character swings his boulder through 40 levels – from the underworld up to Mount Olympus – and the judges deemed it the best.

“It was a defining moment,” Schecter said. He thought, “I want to do this. I want to do this for a living.”

A longtime gamer, Schecter had considered game design as a career, but the computer science major didn’t know how to make that happen until he discovered the Game Design Initiative at Cornell (GDIAC) program. “It literally changed my life,” he said. “It gave me everything I needed to understand how to make a game and also showed me what the real experience was like.”

Founded in 2001, GDIAC was the first undergraduate game design program at an Ivy League school and one of the first in the country. Students can declare game design as a minor, and Princeton Review lists it among the top 50 game design programs for undergrads.

For many GDIAC students, the showcase is when a passion for game design crystallizes into a career plan – all that hard work rewarded by the intoxicating rush of someone entranced by a game you created.

This year, the showcase was held May 20 in Clark Atrium in the Physical Sciences Building, with almost 600 visitors. According to vote counts, the crowd favorites were the desktop game Munchkey, where a monkey chef defeats dangerous fruit to make fruit skewers, and the mobile game Sunk Cost, a multiplayer stealth game where players are either treasure hunters exploring an underwater shipwreck or spirits trying to thwart them. 

“There are a lot of flashy game design programs out there, but we hit above our weight for the resources that we have,” said Walker White, M.S. ’98, Ph.D. ’00, director of GDIAC and senior lecturer in the Department of Computer Science in the Cornell Ann S. Bowers College of Computing and Information Science. 

Despite being a highly competitive field, alumni of the program can be found at all levels of the industry, from AAA companies like PlayStation and Oculus VR, to successful indie teams with breakout hits. Not bad for a minor. 

Level 1: Origin story

GDIAC originated in the mind of David Schwartz, who was hired as a lecturer in 1999, and is now the director of the School of Interactive Games and Media at Rochester Institute of Technology. His background was in civil engineering, but he had written two textbooks on engineering software while completing his dissertation.

Schwartz admits he didn’t know much about computer science when he arrived, but after a colloquium by Don Greenberg, the Jacob Gould Schurman Professor of Computer Graphics, where Greenberg used familiar physics terms to discuss his computer graphics work, Schwartz realized game physics could be a potential area of study. As the faculty advisor for Cornell's Association of Computer Science Undergraduates (ACSU), Schwarz proposed that, instead of playing games, students could learn to design games of their own.

“I thought the students would find this very motivating," Schwartz said. "Just imagine the passion they would have to make these things.” 

While the Department of Computer Science did not yet consider video games to be an academic subject, Charlie van Loan, professor emeritus of computer science and the chair at the time, agreed Schwartz could offer an independent study – but on his own time.

In 2001, Schwartz recruited Rama Hoetzlein ’01, who had dual degrees in computer science and art, and is now an assistant professor of digital media design at Florida Gulf Coast University. He also hired Mohan Rajagopalan M.S. ’02, who later left for the game industry, working on Plants vs Zombies, among other titles. 

The trio founded the program and developed the first classes, making connections in the departments of art, music and communication to help students enhance those aspects in their games. Early funding came from a seed grant from Microsoft, a donation from a trustee, and funds earmarked for Uris Library to design a computer lab, which ultimately became the Cornell Library Collaborative Learning Computer Lab.

Schwartz and Rajagopalan finally got permission to establish game design as a minor in 2007, when only a handful of schools were offering undergraduate courses in game design. “Cornell was actually part of the first wave,” Schwartz said. He estimates that almost one-third of computer science majors at the time took his game design course. 

Level 2: Training grounds

In 2007, White became the program director. He had started designing games as part of a club in his dorm room at Dartmouth, and had recently published research on how database technology could be leveraged to advanced game design, which enhanced the initiative's academic legitimacy. White envisioned GDIAC as an applied software engineering course for sophomores, where the application just happened to be games.

White runs his introductory and advanced game development courses as studio classes, with students constantly presenting and critiquing each other’s games. Twelve teams of eight – a mix of artists, programmers and a musician – function like an indie game company to produce all or part of a game for the annual showcase. Students not only learn sound design, computer graphics, and software engineering but also the critical art of project management. 

While the professional tools for creating games have steadily improved in recent years, the students still build parts of their games from scratch – a decision that White said sets GDIAC apart from other programs. “I'm an educator, and I believe in teaching foundational, portable skills,” White said. “I want them seeing how everything fits together.”

In addition to the showcase, students have frequently competed in juried festivals, such as the Boston Festival of Indie Games, but mainly before the COVID-19 pandemic. The program’s biggest breakout hit was the mobile game Family Style – a party game where players pass ingredients between phones to assemble dishes. In 2019, the game was featured on the front page of the Apple store, went viral in Thailand, and reached a peak of 15,000 daily active players, White said.

But despite the preparation GDIAC provides, game development is still a tough field, with stiff competition for jobs, grinding hours, and frequent layoffs. “I love supporting my students who want to go into the game industry, but I don't sugarcoat it,” White said. He sends five or six graduates off to the game industry each year, while 15 to 20 others go on to work in game-related tech. “I make sure people are going in wide-eyed, because they're hopefully in it for the long haul.”

Level 3: Achievements unlocked

Developing video games may seem like child’s play, but it’s a big business. The Entertainment Software Association reports the U.S. spent $56.6 billion on video games in 2022; worldwide, the market was estimated to be worth $214.2 billion. About two-thirds of people in the U.S. play video games at least once a week, and the industry employed more than 400,000 people in 2020.

“It's definitely a challenging industry; the bar is generally really high,” said Rajiv Puvvada '10, an early graduate of the program and industry veteran who has worked at top companies, including Zynga, Juicebox Games, and AWS for Games. Especially for entry-level positions, “there are only so many of those spots that come around,” he said.

Puvvada can’t remember a time when he wasn’t gaming; there are photos of him reaching for a Nintendo controller from his high chair. He started making games as a teenager, “going about things in the entirely wrong way,” but he felt the stigma that game development wasn’t a “real” job. Through GDIAC, he discovered a path to the career he wanted. 

Working in a multi-disciplinary team with deadlines at Cornell was excellent preparation for the real thing, he said, and learning the fundamentals has helped him stay agile as gaming technology evolves. “The game industry field is very wide, and it changes a lot, often very quickly,” he said.

Puvvada said he’s always happy to talk with current students and help them enter the industry and grow their career. He thinks this is vital for “increasing the level of diversity in the industry, because that's something that benefits all of us.”

There are no cheat codes for breaking into the industry, but recent graduate Kristina Gu '21, M.Eng '22, said after multiple attempts, she felt fortunate to have landed an internship at PlayStation. Thanks to a good word put in by her internship supervisor, Gu is now a technical designer at Naughty Dog, an affiliated company behind the critically acclaimed game The Last of Us.

GDIAC made her well-rounded, which makes her a more flexible game designer, she said. “Game development is really one of the only fields that has that perfect balance of being interdisciplinary, technical and creative, which is something I really, really enjoy.” 

A long-time gamer, Gu became captivated by game design while playing Uncharted 4, another Naughty Dog title, in high school. She’s been a rabid fan since.

Even Gu’s parents are largely happy with where she’s at. “My dad is still holding out hope that maybe I’ll go back to business school, but I think they’re pretty happy with how it’s gone,” she said. 

The End: And another quest begins

After graduation, Schecter will be the latest GDIAC alum to join the game industry, with a position developing graphics at Blizzard Entertainment. He interned at the company in 2022 after being selected from a pool of about 20,000-30,000 applicants.

“I love Blizzard games. I actually run the Overwatch team at Esports at Cornell, which is a Blizzard title,” he said. “I’m sure that helped.”

Schecter will once again descend to the underworld to work on Diablo 4, the fourth installment of their popular Diablo series – a set of dungeon crawler games, where players contend with hordes of demons.

Soon, Schecter’s games will not only be played by hundreds of people at the game design showcase, but will captivate millions of people around the world.

Patricia Waldron is a writer for the Cornell Ann S. Bowers College of Computing and Information Science.