The Allen School has established a new center at the University of Washington that aims to catalyze the next generation of cloud computing technology. The Center for the Future of Cloud Infrastructure, or FOCI, will cultivate stronger partnerships between academia and industry to enable cloud-based systems to reach new heights when it comes to security, reliability, performance, and sustainability.
“The first generation of the cloud disrupted conventional computing but focused on similar engineering abstractions, which is typical of many new technologies,” said Allen School professor Ratul Mahajan, co-director of the FOCI Center and, until recently, co-founder and CEO of cloud computing startup Intentionet. “Now that cloud computing is on the cusp of a more radical transformation, this center will help usher in a new era by cultivating tighter partnerships between researchers and practitioners to address emerging bottlenecks and explore new opportunities.”
That transformation is being driven in large part by the rise in machine learning, edge computing, 5G and other burgeoning technologies. According to Mahajan’s Allen School colleague and center co-director Simon Peter, the demands of these new workloads — including exponential growth in the energy required to power their applications — will require researchers to rethink the full computing stack from the ground up.
“Companies and consumers are seeking ever-greater levels of security, reliability and performance in the cloud at a reduced cost,” Peter noted. “Not just monetary cost, but also in terms of cost to the environment. For a while, thanks to Moore’s Law, we were gaining ground when it comes to energy efficiency. But now the gains have slowed or even reversed; for example, in the U.S. the energy demand for computation is growing twice as fast as solar and wind power. So we need to think holistically about the hardware-software interface and how to make cloud computing sustainable as well as resilient and secure.”
One of the areas that Peter and his colleagues are keen to explore is energy-aware cloud computing, which would enable tradeoffs between power and performance while making cloud applications resilient to disruption. Another potential avenue of inquiry concerns how the development of systems to effectively manage the variety of hardware accelerators used in settings such as disaggregated storage and emerging machine learning applications while minimizing latency, ensuring fairness, and meeting multi-dimensional resource needs — among other challenges.
The co-directors of the new FOCI Center at the UW, top, from left: Arvind Krishnamurthy and Ratul Mahajan; bottom left: Simon Peter
How the center approaches these challenges will be informed by a technical advisory board comprising representatives of cloud companies Alibaba, Cisco, Google, Microsoft and VMware — all significant movers and shakers in the cloud space. Their input will help guide the center’s research toward real-world impact based on current trends, what problems they anticipate over a five to 10- year time horizon, and how solutions might be applied in practice. Center researchers will apply these practical insights to their pursuit of big, open-ended ideas, drawing upon cross-campus expertise in systems, computer architecture, networking, machine learning, data science, security, and more.
“Industry knows the pain points and technology trends; academia is adept at the exploratory, collaborative work that’s fundamental to solving hard problems,” noted Allen School professor and center co-director Arvind Krishnamurthy, who also serves as an advisor to UW machine learning spinout OctoML. “By bringing the two together, this center will not only yield compelling solutions but also contribute to the education of students who will go on to build these next-generation systems.”
The FOCI Center was seeded with industry commitments totaling $3.75 million over three years. The Allen School is hosting a launch event on the UW campus in Seattle today to connect faculty and student researchers with industry leaders interested in shaping the future of cloud computing.
“Seattle is the cloud city, both in weather and as home to the largest cloud companies, so it was only natural to establish a center focused on cloud computing and leverage the synergies between the UW’s research expertise and our local industry leadership of this space,” said Magdalena Balazinska, professor and director of the Allen School. ”When it comes to what we can accomplish together, I would say the sky’s the limit.”
Yejin Choi (John D. and Catherine T. MacArthur Foundation)
Yejin Choi, a professor in the Allen School’s Natural Language Processing group, was selected as a 2022 MacArthur Fellow by the John D. and Catherine T. MacArthur Foundation to advance her work “using natural language processing to develop artificial intelligence systems that can understand language and make inferences about the world.” The MacArthur Fellowship — also known as the “genius grant” — celebrates and invests in talented and creative individuals whose past achievements signify their potential to make important future contributions. Each recipient receives a stipend of $800,000 that comes with no strings attached.
“It’s been several weeks since I learned about this award, and it still feels so surreal,” Choi told UW News.
Choi, who joined the Allen School faculty in 2014, may feel like she is dreaming, but her work has had a very real impact. Currently the Brett Helsel Career Development Professor in the Allen School and senior research manager for the Mosaic team at the Allen Institute for AI, Choi has contributed to a series of high-profile projects that have expanded the capabilities of natural language models — and uncovered potential pitfalls. For example, she was among the first to bridge the fields of NLP and computer vision by teaching models to generate original and accurate image descriptions based on visual content in place of conventional statistical approaches. She has also contributed to a variety of tools for analyzing and combating the proliferation of bias and misinformation online, from AI-generated “fake news” to trashy training inputs that lead to toxic language degeneration, along with new methods for assessing the quality of open-ended machine-generated text compared to that generated by humans.
Choi and her collaborators went a step further with the development of Ask Delphi, an experimental platform for exploring how machines might acquire and exercise moral judgment in response to real-world situations. Through this and other work, Choi is pushing the field closer to her overarching goal: to imbue machines with a human-like ability to reason and communicate about the world in both physical and abstract terms. Whatever comes next, Choi is determined to fulfill the spirit of the Fellowship by pursuing the most original and impactful research ideas — even when they are accompanied by a degree of risk.
“Taking the road less traveled may seem exciting at first, but sustaining this path can be lonely, riddled with numerous roadblocks and disheartening at times,” Choi said. “This fellowship will power me up to go ahead and take that adventurous route.”
Previous MacArthur Fellowship winners with an Allen School connection include Choi’s faculty colleague Shwetak Patel, alumni Stefan Savage (Ph.D., ‘02), a professor at the University of California San Diego, and Christopher Ré (Ph.D., ‘09), a professor at Stanford University, and former Allen School faculty member Yoky Matsuoka, currently founder and CEO of Yohana.
A sea of student backpacks stashed outside the October 4 career fair
After several years of Covid-induced online career fairs, the Allen School returned to an in-person format this fall!
On October 4 and 6, more than 50 companies — members of the Allen School’s Industry Affiliates program — came to campus to recruit students for full-time, part-time, and internship positions. On each day, the first half of the session was devoted to Allen School students; UW students in related majors joined for the second half. In all, more than 1,000 students participated across the two days.
The fall career fairs are instrumental in connecting students with career opportunities in the local technology community. In 2021-22, the Allen School alone sent more than 60 graduates to full-time positions at Amazon, more than 50 to Google, and more than 40 each to Facebook and Microsoft. Allen School students accepted full-time positions at more than 100 tech companies in total — the vast majority in the Puget Sound region.
When bees leave the hive, they can spend all day flying and foraging on a single “charge” owing to their ability to convert fats and carbohydrates that store significantly more energy than batteries. When other insects traverse the landscape, the structure of their retinas combined with the motion of their heads enable them to efficiently take in and process visual information. And when dandelions shed their seeds, structural variations ensure that they are dispersed through the air over short and long distances to cover maximum ground.
Allen School professor Vikram Iyer is not a biologist, but he takes inspiration from these and other biological phenomena to engineer programmable systems and devices that can go where computers have been unable to go before — and solve problems more efficiently and safely than previously thought possible. During his time as a University of Washington Ph.D. student, Iyer imagined how the so-called internet of biological and bio-inspired things could transform domains ranging from agriculture to wildlife conservation. His results recently inspired the Association for Computing Machinery’s Special Interest Group on Mobility of Systems, Users, Data, and Computing to recognize him with the SIGMOBILE Doctoral Dissertation Award for “creative and inspiring work that shows how low-power sensing, computing and communication technologies can be used to emulate naturally-occurring biological capabilities.”
“Building bio-inspired networking and sensing systems requires expertise across multiple disciplines spanning computer science, electrical engineering, mechanical engineering and biology,” said Allen School professor Shyam Gollakota, Iyer’s Ph.D. advisor. “I think this thesis breaks new ground by designing programmable technologies that not only mimic nature but also takes the crucial step of integrating electronics with living organisms.”
Gollakota and Iyer worked together on a series of projects that gave new meaning to the term “computer bug” — but in this case, they took their lessons from the kind of bugs with legs and wings. For one of their early projects, Living IoT, the team developed a scaled down wireless sensing and communication platform that was light enough to be worn by bumblebees in flight. The tiny sensor backpacks incorporate a rechargeable power source, localization hardware, and backscatter communication to relay data once the bee returns to the hive in a form factor topping out at a mere 102 micrograms. Later, when the northern giant hornet — colloquially referred to as the “murder hornet” — was sighted in northwest Washington, the state’s Department of Agriculture enlisted Iyer’s help in designing and affixing tiny tracking devices onto a live specimen so that agency staff could track it back to the nest.
After seeing their concept of on-board sensors for insects take flight, Iyer and his colleagues came back down to earth to develop a tiny wireless camera inspired by insect vision. Their system, which they dubbed “Beetlecam,” offered a fully wireless, autonomously powered, mechanically steerable vision system that imitates the head motion of insects. By affixing the camera onto a moveable mechanical arm, the team could mimic insect head motion to capture a wide-angle view of the scene and track the movement of objects while expending less energy — and at a higher resolution. The complete vision system, which can be controlled via smartphone, is small enough to mount on the back of a live beetle or insect-sized terrestrial robots such as their own prototype built to demonstrate the capabilities.
Many sensors, including those designed by Iyer and his collaborators, still require a method of transportation, be it beetle, bee, or robot. Iyer and his collaborators wondered if they could design sensors capable of delivering themselves. The answer, to borrow a phrase from singer/songwriter Bob Dylan, was blowing in the wind — in the form of dandelion seeds. Iyer and the team developed a wireless, solar-powered sensing and communication system that can be carried aboard flexible, thin shapes. The shapes are designed to carry the sensors through air and land upright 95% of the time, relying on a structure reminiscent of the bristle-like shape of dandelion seeds — with some necessary modifications to accommodate the weight of the attached sensor. The team also demonstrated that, by modulating the porosity and diameter of the structures, they can ensure the sensors are dispersed at various distances like the seeds.
Unlike many miniaturized systems, Iyer’s flora- and fauna-inspired projects favor designs that rely on off-the-shelf parts instead of requiring custom-built circuits.
“In addition to showing how we can take lessons from nature to advance a new category of bioinspired computing, my work demonstrates how we can use programmable general-purpose components to rapidly develop these novel miniaturized wireless systems,” Iyer explained. “This approach has the potential to exponentially increase innovation in domains such as smart agriculture, biological tracking, microrobots, and implantable devices. My goal is to enable anyone with a computer engineering background to advance miniaturized systems without the need to also develop custom silicon.”
Iyer, who earned his Ph.D. from the UW Department of Electrical & Computer Engineering before joining the Allen School faculty last year, previously earned a Paul Baran Young Scholar Award from the Marconi Society and his work was voted Innovation of the Year in the 2021 GeekWire Awards. He is the third student researcher advised by Gollakota to win this award in recent years, following in the footsteps of Allen School alum Rajalakshmi Nandakumar (Ph.D., ‘20), now a faculty member at Cornell University, and ECE alum Vamsi Talla (Ph.D., ‘16), who was co-advised by Allen School and ECE professor Joshua Smith and is currently CTO of UW spinout Jeeva Wireless.
Congratulations, Vikram!
Photo credit: Mark Stone/University of WashingtonRead more →
Maya Cakmak beside the plaque dedicated to computer scientist and visionary Anita Borg in the Bill & Melinda Gates Center for Computer Science & Engineering on the University of Washington’s Seattle campus. Borg began her undergraduate education at the UW. Photo courtesy of Maya Cakmak
For Allen School professor Maya Cakmak, the future of robotics hinges on the human element. Since the early days of her research career, Cakmak has been leveraging advances in human-computer interaction and accessibility to shift robotics research from primarily technology-centric approaches toward a more user-centric approach. She is also known for putting people first through her support for programs and policies aimed at increasing participation in computing by women and people with disabilities. For her efforts, the Computing Research Association’s Committee on Widening Participation in Computing Research (CRA-WP) recently recognized Cakmak with its 2022 Anita Borg Early Career Award — named in honor of another woman in computing who wasn’t afraid to break new ground and lead by example.
Cakmak holds the Robert E. Dinning Career Development Professorship in the Allen School, where she directs the Human-Centered Robotics Lab. There, she and her collaborators develop robots that can be programmed by users with diverse needs and preferences to assist with everyday tasks — in her own words, “empowering users to decide what their robots will do for them.”
Even before she joined the University of Washington faculty in 2013, Cakmak had already begun building a reputation in robotics circles for her human-centric approach. As a Ph.D. student at Georgia Tech, Cakmak showed that many data-driven methods for programming by demonstration were ill-suited to novice users, proposing instead to employ interaction-driven approaches that incorporate user studies — a novel idea at the time that has since entered the mainstream. She also was an early evangelist of active learning in robotics, which enables robots to acquire new knowledge and skills via queries instead of just passively receiving data from humans.
“I deeply care about the relevance and usefulness of my research.” Cakmak presenting at the Robotics: Science and Systems conference in 2018. Anca Dragan
After her arrival at the Allen School in 2013, Cakmak sought to expand end-user programming capabilities for robots beyond conventional programming by demonstration by drawing upon techniques from HCI. She also pursued new connections between robotics and programming languages, uniting two of the Allen School’s strengths to usher in a new and growing area of interdisciplinary research. For example, Cakmak and Ph.D. student Sonya Alexandrova teamed up with faculty colleague Zachary Tatlock in the Programming Languages & Software Engineering (PLSE) group on RoboFlow, an intuitive visual programming language to enable users to program robots to perform mobile manipulation tasks in various real-world environments. She also worked with then-student Justin Huang (Ph.D., ‘18) and Allen School alum Tessa Lau (Ph.D., ‘01), who at the time served as CTO and “Chief Robot Whisperer” at Savioke, Inc., to create a rapid programming system for mobile service robots. That system, which the team dubbed CustomPrograms, featured a cloud-based graphical interface to support the rapid prototyping and development of custom applications by experts as well as inexperienced programmers.
Cakmak and Huang followed that up with Code3, a suite of user-friendly tools for perception, manipulation and high-level programming that enables rapid, end-to-end programming of robots such as the PR2 and Fetch to perform mobile manipulation tasks. As with many of her projects, Code3 was designed to appeal to experts and novice users alike. In 2016, Cakmak received a CAREER Award from the National Science Foundation for her project titled “End-user programming of general-purpose robots” to continue this line of research.
According to her Allen School robotics colleague Dieter Fox, Cakmak’s human-centric approach has been hugely influential within the robotics community.
”Human-robot interaction has grown tremendously over the past decade, and Maya has been at the forefront of the field,” said Fox, director of the UW’s Robotics and State Estimation Lab and senior director of robotics research at NVIDIA. “Her work has also had real-world impact through her collaborations with multiple robotics companies that ship their robots with end-user tools that she developed.”
Fox knows all about the potential impact of such academic-industry collaboration. He and his colleagues at NVIDIA recently teamed up with Cakmak to develop new capabilities in human-robot handovers — an essential skill for robots to safely and reliably assist humans with everyday tasks. The team’s system, which demonstrated smooth human-to-robot handover of arbitrary objects commonly found in people’s homes, earned the Best Paper Award in HRI at the International Conference on Robotics and Automation (ICRA 2021). Cakmak also collaborated with Anat Caspi, director of the Allen School’s Taskar Center for Accessible Technology, and researchers in the Personal Robotics Lab led by professor Sidd Srinivasa to explore user preferences and community-centered design principles to inform the development of robot-assisted feeding systems for people with mobility impairments.
Cakmak (right) introduces participants in the Allen School’s K-12 outreach program to Rosie the robot and programming by demonstration. Photo courtesy of Maya Cakmak
Whatever problem she aims to solve, Cakmak has been eager to take her cue from the people who have the potential to benefit most from her work.
“I deeply care about the relevance and usefulness of my research,” Cakmak explained in her Early Career Spotlight talk at the 2018 Robotics: Science & Systems (RSS) conference. “To that end, I try to evaluate systems I develop with a realistic and diverse set of tasks, I put these systems in front of real potential users with diverse backgrounds and abilities, and I take every opportunity to demonstrate and deploy them in the real world.”
Cakmak’s contributions to accessibility go far beyond her projects and papers; she is also a vocal proponent of making the act of research — and the conferences and other events where research careers are made — more inclusive of people with disabilities and women. This includes making it easier for women who are new mothers to participate, inspired by her personal experience. In one specific example, Cakmak successfully lobbied for changes in one organization’s reimbursement policies to cover childcare expenses for invited speakers.
“When I had my first child, I received a transitional grant from the UW ADVANCE Center for Institutional Change to cover expenses of taking my infant along for work travel,” Cakmak said. “Continuing to go to conferences was critical for me to stay active in my research community and help my graduate students network. But it did more than that. Colleagues would often approach me, not just to play with the baby but to ask questions about the logistics of taking your baby to a conference or how to manage starting a family while being on tenure-track. Many were amazed to hear about the ADVANCE grant and went back to ask for similar initiatives at their institutions.
“In an academic career system where young parents are disadvantaged, we are learning how to make it work from one another, while also pushing for positive institutional change,” she continued. “This is why representation matters so much. We owe many privileges we now take for granted to the hard work of those who were disadvantaged back in the day.”
Back on campus, Cakmak served as a co-principal investigator on the National Science Foundation-funded AccessEngineering initiative to incorporate topics such as universal design into engineering courses and to make labs and maker spaces accessible to students with diverse abilities. In an effort to engage more people with disabilities in computing and, in particular, robotics, Cakmak designed and taught a course for high school students over multiple summers as part of the UW DO-IT Center’s Scholars program. The course provided participants with hands-on experience in programming robots while encouraging them to think about how the field can help solve real-world problems. She developed a similar workshop for AccessComputing’s OurCS@UW program for undergraduate women with disabilities. Cakmak’s contributions inspired the DO-IT Center to recognize her with its 2021 Trailblazer Award.
Cakmak has been a vocal proponent of policies that enable young parents to continue to attend research conferences and advance their careers. Aditya Mandalika
In addition to these activities, Cakmak developed a seminar to provide second-year students with the foundational skills to participate in research. She also served as the UW faculty representative for the fourth cohort of the LEAP Alliance, a program that seeks to diversify the professoriate by supporting underrepresented students in pursuing academic careers. She also organized a rejuvenated RSS Women in Robotics Workshop, including raising funds for travel grants that enabled women roboticists from around the world to participate, and later supported its expansion to include researchers from other underrepresented groups.
“Maya is a stellar researcher who is a leader in her field of research, and she has dedicated an immense amount of time and effort to broadening participation by women and students with disabilities — directly and indirectly impacting many students,” Fox said. “I cannot imagine a more deserving recipient of the Anita Borg Award.”
The award includes a $5,000 prize, which Cakmak donated to the Allen School to support more women to pursue undergraduate research.
“This award means so much to me because it recognizes things I did in my career beyond my research that I did not expect to be recognized for,” Cakmak said. “I have been so fortunate to have many amazing role models, mentors, advocates, and supporters; I was just trying to pay it forward. I am honored especially because I have been so inspired by Anita Borg.”
Indeed, Cakmak’s work embodies Borg’s famous quote: “If we want technology to serve society rather than enslave it, we have to build systems accessible to all people — be they male or female, young, old, disabled, computer wizards or technophobes.”
Cakmak is the second Allen School faculty member to earn the Borg Early Career Award in the past five years, following professor Yejin Choi’s recognition in 2018. Allen School alumni Martha Kim (Ph.D., ’08), a faculty member at Columbia University; A.J. Bernheim Brush (Ph.D., ’02), Partner Group Program Manager at Microsoft; and Gail Murphy (Ph.D., ’96), Vice President of Research and Innovation and a faculty member at the University of British Columbia, are also past recipients of the award.
For millions of people who participate in activities such as snorkeling and scuba diving each year, hand signals are the only option for communicating safety and directional information underwater. While recreational divers may employ around 20 signals, professional divers’ vocabulary can exceed 200 signals on topics ranging from oxygen level, to the proximity of aquatic species, to the performance of cooperative tasks.
The visual nature of these hand signals limits their effectiveness at distance and in low visibility. Two-way text messaging is a potential alternative, but one that requires expensive custom hardware that is not widely available.
Researchers at the University of Washington show how to achieve underwater messaging on billions of existing smartphones and smartwatches using only software. The team developed AquaApp, the first mobile app for acoustic-based communication and networking underwater that can be used with existing devices such as smartphones and smartwatches.
“Smartphones rely on radio signals like WiFi and Bluetooth for wireless communication. Those don’t propagate well underwater, but acoustic signals do,” said co-lead author Tuochao Chen, a Ph.D. student in the Allen School. “With AquaApp, we demonstrate underwater messaging using the speaker and microphone widely available on smartphones and watches. Other than downloading an app to their phone, the only thing people will need is a waterproof phone case rated for the depth of their dive.”
The AquaApp interface enables users to select from a list of 240 pre-set messages that correspond to hand signals employed by professional divers, with the 20 most common signals prominently displayed for easy access. Users can also filter messages according to eight categories, including directional indicators, environmental factors, and equipment status.
In building the app, Chen and his collaborators — co-lead author and fellow Ph.D. student Justin Chan and professor Shyam Gollakota — had to overcome a variety of technical challenges that they haven’t previously encountered on dry land.
“The underwater scenario surfaces new problems compared to applications over the air,” explained Chan. “For example, fluctuations in signal strength are aggravated due to reflections from the surface, floor and coastline. Motion caused by nearby humans, waves and objects can interfere with data transmission. Further, microphones and speakers have different characteristics across smartphone models. We had to adapt in real time to these and other factors to ensure AquaApp would work under real-world conditions.”
Those other factors include the tendency for devices to rapidly shift position and proximity in the current and the various noise profiles the app might encounter due to the presence of vessels, animals, and even low-flying aircraft.
The team created an algorithm that allows AquaApp to optimize, in real time, the bitrate and acoustic frequencies of each transmission based on certain parameters, including distance, noise and variations in frequency response across devices. When one user wants to send a message to another device, their app first sends a quick note, called a preamble, to the other device. AquaApp on the second device runs the algorithm to determine the best conditions to receive the preamble; it then tells the first device to use those same conditions to send the actual message.
The researchers developed a networking protocol to share access to the underwater network, akin to how WiFi networks referee internet traffic, to support messaging between multiple devices. AquaApp can accommodate up to 60 unique users on its local network at one time.
The team tested the real-world utility of the AquaApp system in half a dozen locations offering a variety of water conditions and activity levels, including under a bridge in calm water, at a popular waterfront park with strong currents, next to the fishing dock of a busy lake, and in a bay with strong waves. In a series of experiments, they evaluated AquaApp’s performance at distances of up to 113 meters and depths of up to 12 meters.
“Based on our experiments, up to 30 meters is the ideal range for sending and receiving messages underwater, and 100 meters for transmitting SoS beacons,” Chen said. “These capabilities should be sufficient for most recreational and professional scenarios.”
The researchers also measured AquaApp’s impact on battery life by continuously running the system on two Samsung Galaxy S9 smartphones at maximum volume and with screens activated. The app reduced the devices’ battery power by just 32% over the course of four hours, which is within the maximum recommended dive time for recreational scuba diving.
“AquaApp brings underwater communication to the masses,” said Gollakota, who directs the Mobile Intelligence Lab and holds the Torode Family Career Development Professorship in the Allen School. “The state of underwater networking today is similar to ARPANET, the precursor of the internet, in the 1970s, where only a select few had access to the internet. AquaApp has the potential to change that status quo by democratizing underwater technology and making it as easy as downloading software on your smartphone.”
In the spring of 2020, people took to the streets — and to the tweets — in protest after a white police officer murdered George Floyd, a Black man, in Minneapolis by kneeling on his neck and back for over nine minutes. Black Lives Matter, a movement spawned seven years earlier following the shooting death of Trayvon Martin, an unarmed Black teenager, in Florida and the killer’s subsequent acquittal, emerged as the online and offline rallying cry against police brutality and racist violence perpetrated against Black people.
Following Floyd’s death, Twitter users flooded the social media platform with hundreds of millions of posts expressing a range of emotions concerning Black Lives Matter and the campaign for racial justice. A team led by Allen School professor Yulia Tsvetkov sought to facilitate a greater understanding of the connection between those emotional expressions and the narrative surrounding Black Lives Matter and its supporters — and how that connection could also shed light on the role of social media messaging in online activism and large-scale social movements, more generally. To that end, the researchers applied recent advances in natural language processing to analyze the content of 34 million original English-language tweets about BLM posted in 2020 between May 24 and June 28 to identify the prevailing emotions expressed on social media about the movement and associated protests.
“While we identified high levels of anger and disgust across all posts in our dataset, what jumped out at us was the prevalence of positive emotions in posts containing pro-BLM hashtags such as #BlackLivesMatter, #JusticeforFloyd and #NoJusticeNoPeace, and correlating with on-the-ground protests,” said co-lead author and Stanford University postdoc Anjalie Field, who worked on the project as a visiting UW student and Ph.D. candidate at Carnegie Mellon University. “Positive emotions like hope and optimism are more prevalent in posts with explicitly pro-BLM hashtags than other subsets of the data, which contradicts the stereotype of BLM supporters as promoting anger and outrage.”
To perform their analysis, Field, Tsvetkov and their collaborators developed a neural classification model based on the six core emotions identified by psychologist Paul Ekman: anger, disgust, fear, surprise, positivity — what Ekman refers to as “joy” — and sadness. Consistent with prior work applying Ekman’s taxonomy, the researchers treated each category as a superset of finer grained emotions such as the aforementioned hope and optimism (positivity), disapproval and rage (anger), vigilance and apprehension (fear), confusion and curiosity (surprise), and grief and remorse (sadness). Because neural models trained on a pre-collected data set may perform poorly when used to infer meaning from text gathered in a different domain, the team opted for a domain adaptation approach. By combining task-adaptive pre-training with few-shot learning techniques, their methodology permits the re-use of existing annotated datasets to train the model on different domains, rather than going through the effort and expense of collecting and annotating new datasets. They trained their model on two large, pre-existing social media datasets annotated for expressed emotions and adapted it to tweets relating to Black Lives Matter.
As Field noted, anger was the prevailing emotion identified by the model, which detected its presence in 40% or more of the tweets posted over the course of the month. Disgust and positivity alternated as the second-most prevalent emotions, with surprise, sadness and fear rarely approaching 10%. When the researchers analyzed the subset of 6.5 million tweets containing pro-BLM hashtags, they found that positivity consistently outweighed the other categories starting around four days in — on some days, more than double the negative emotions.
Drilling down into the dataset according to date enabled the team to identify instances where emotions spiked, presumably in connection with events. For example, anger and sadness peaked in tweets with pro-BLM hashtags in the days following Floyd’s death and prior to the first weekend of protests. Positivity, meanwhile, rose in the days leading up to that weekend and afterward became the most frequently expressed emotion through the rest of the month. Positivity peaked on the Juneteenth holiday, present in 60% of tweets carrying pro-BLM hashtags. By analyzing location data, the researchers also found that the volume of tweets expressing positive emotions positively correlated with users’ proximity to on-the-ground protests.
Such rich analysis from the team’s adaptive neural model contrasts with that of conventional social science analyses, which typically rely on more rigid lexicon-based approaches using a set list of words associated with an emotion to determine whether content reflects that emotion.
“Lexicon-based models are easy to use, but they aren’t particularly good at capturing broader connotations or adapting to new contexts,” explained co-lead author Chan Young Park, a visiting UW student who is pursuing a Ph.D. from CMU. “For example, one popular model connotes the word ‘police’ with the terms ‘fear,’ ‘positive’ and ‘trust’ — emotions that are unlikely to factor into protests against police brutality. Our framework offers a more robust method for extracting accurate social meaning from text data that can be adapted to different contexts and language varieties.”
The meaning the team extracted from posts using its computational model is consistent with findings from previous social psychology studies; while moral outrage and anger can prompt people to become involved in social movements, positive emotions are necessary to sustain that involvement over time.
“Words and emotions are powerful tools for online activism. Emerging NLP techniques are also powerful tools that can help us understand how those words and emotions contribute to building and sustaining social movements,” said Tsvetkov. “In doing so, they can help us also to dispel negative stereotypes about marginalized communities that lead to physical, social and economic harm.”
Additional contributors to the paper include co-lead author and then-student Antonio Theophilo, who recently earned his Ph.D. from the Institute of Computing at the University of Campinas in Brazil, and Jamelle Watson-Daniels, a Ph.D. student at Harvard University and Director of Research at Data for Black Lives.
For people around the world, technology eases the friction of everyday life: bills paid with a few clicks online, plans made and sometimes broken with the tap of a few keys, professional and social relationships initiated and sustained from anywhere at the touch of a button. But not everyone experiences technology in a positive way, because technology — including built-in safeguards for protecting privacy and security — isn’t designed with everyone in mind. In some cases, the technology community’s tendency to develop for a “default persona” can lead to harm. This is especially true for people who, whether due to age, ability, identity, socioeconomic status, power dynamics or some combination thereof, are vulnerable to exploitation and/or marginalized in society.
Researchers in the Allen School’s Security & Privacy Research Lab have partnered with colleagues at the University of Florida and Indiana University to provide a framework for moving technology design beyond the default when it comes to user security and privacy. With a $7.5 million grant from the National Science Foundation through its Secure and Trustworthy Cyberspace (SaTC) Frontiers program, the team will blend computing and the social sciences to develop a holistic and equitable approach to technology design that addresses the unique needs of users who are underserved by current security and privacy practices.
“Technology is an essential tool, sometimes even a lifeline, for individuals and communities. But too often the needs of marginalized and vulnerable people are excluded from conversations around how to design technology for safety and security,” said Allen School professor and co-principal investigator Franziska Roesner. “Our goal is to fundamentally change how our field approaches this question to center the voices of marginalized and vulnerable people, and the unique security and privacy threats that they face, and to make this the norm in future technology design.”
To this end, Roesner and her collaborators — including Allen School colleague and co-PI Tadayoshi Kohno — will develop new security and privacy design principles that focus on mitigating harm while enhancing the benefits of technology for marginalized and vulnerable populations. These populations are particularly susceptible to threats to their privacy, security and even physical safety through their use of technology: children and teenagers, LGBTQ+ people, gig and persona workers, people with sensory impairments, people who are incarcerated or under community supervision, and people with low socioeconomic status. The team will tackle the problem using a three-prong approach, starting with an evaluation of how these users have been underserved by security and privacy solutions in the past. They will then examine how these users interact with technology, identifying both threats and benefits. Finally, the researchers will synthesize what they learned to systematize design principles that can be applied to the development of emerging technologies, such as mixed reality and smart city technologies, to ensure they meet the privacy and security needs of such users.
The researchers have no intention of imposing solutions on marginalized and vulnerable communities; a core tenet of their proposal is direct consultation and collaboration with affected people throughout the duration of the project. They will accomplish this through both quantitative and qualitative research that directly engages communities in identifying their unique challenges and needs and evaluating proposed solutions. The team will apply these insights as it explores how to leverage or even reimagine technologies to address those challenges and needs while adhering to overarching security and privacy goals around the protection of people, systems, and data.
The team’s approach is geared to ensuring that the outcomes are relevant as well as grounded in rigorous scientific theory. It’s a methodology that Roesner, Kohno, and their colleagues hope will become ingrained in the privacy and security community’s approach to new technologies — but they anticipate the impact will extend far beyond their field.
Tadayoshi Kohno (left) and Franziska Roesner. Dennis Wise
“In addition to what this will mean in terms of a more inclusive approach to designing for security and privacy, one of the aspects that I’m particularly excited about is the potential to build a community of researchers and practitioners who will ensure that the needs of marginalized and vulnerable users will be met over the long term,” said Kohno. “Our work will not only inform technology design, but also education and government policy. The impact will be felt not only in the research and development community but also society at large.”
Kohno and Roesner are joined in this work by PI Kevin Butler and co-PIs Eakta Jain and Patrick Traynor at the University of Florida, co-PIs Kurt Hugenberg and Apu Kapadia at Indiana University, and Elissa Redmiles, CEO & Principal Researcher at Human Computing Associates. The team’s proposal, “Securing the Future of Computing for Marginalized and Vulnerable Populations,” is one of three projects selected by NSF in its latest round of SaTC Frontiers awards worth a combined $24.5 million. The other projects focus on securing the open-source software supply chain and extending the “trusted execution environment” principle to secure computation in the cloud.
Isaiah Lemmon (center) accepting his Dean’s Medal certificate from Dean Nancy Allbritton (left) and chemical engineering professor Jim Pfaendtner (right). Greg DeBow
After graduating from the University of Washington in December with degrees in computer science and chemical engineering, Allen School alum Isaiah Lemmon (B.S., ‘21) took on a software engineering role at Amazon Web Services. There, he intends to put his education to work advancing energy efficient solutions for the datacenter, inspired in part by his experience as an undergraduate researcher during his time at UW. That experience, combined with his rigorous coursework and achievements inside and outside of the classroom, recently earned him a 2022 Dean’s Medal for Academic Excellence from the College of Engineering — affirming his decision to pursue research across multiple disciplines to prepare himself to make a positive difference in the world.
“I’m honored to have been chosen for this award, and I’m really excited about the opportunity the combination of these two degrees has opened to me,” Lemmon said. “The amount of applicable knowledge and skills that different engineering disciplines have to offer each other is incredible, and I’d strongly encourage other students to pursue interdisciplinary research for that exposure.”
From his early days as an undergraduate student, Lemmon displayed a talent and passion for research and discovery. He applied both to great effect working with professor Jim Pfaendtner, chair of the UW Department of Chemical Engineering, on projects that enabled him to explore his dual interests in molecular science and computation. One such project involved studying interactions between titanium dioxide and water at the solid/liquid interface via ab initio molecular dynamics, for which Lemmon performed programming and data analysis using the UW’s HYAK supercomputing cluster.
“Isaiah approached me when he was fresh out of high school, and I quickly realized that he was an incredibly rare student,” said Pfaendtner. “Before long, Isaiah was a regular in our research group and made impressive gains on several projects related to catalysis, interfacial phenomena and ab initio molecular dynamics. However, it was in his work as a de facto software engineer for my team, during his third year of study, that we really saw Isaiah shine.”
At that time, Pfaendtner was eager to keep Lemmon engaged with his lab by supporting him in his quest to combine his knowledge and skills from both majors. He decided to hire Lemmon to assist lab members with turning their prototype software into usable products that could be deployed in the real world. In one instance, Lemmon rewrote a complex piece of reaction engineering software in Python from scratch to make it more robust and useful for a broader range of projects in the lab. He subsequently collaborated with then-Ph.D. student Sarah Alamdari to make modifications to a community software package that enabled Pfaendtner’s group to make their algorithms more widely available within the research community. Lemmon’s mentor characterized this work as evidence that he had already achieved a level of ability and independence more akin to that of a fourth or fifth year Ph.D. student than an undergraduate.
Lemmon also did not shy away from tackling advanced coursework, including material that would typically be the preserve of graduate students. For example, he excelled in Pfaendtner’s Mass Transfer and Separations course and Allen School professor Tom Anderson’s Distributed Systems course — two offerings considered by students to be among the most difficult and intellectually challenging within their respective majors. In his final quarter at the UW, Lemmon enrolled in a new and similarly challenging Datacenter Systems course that explores the technologies that go into the construction and operation of large-scale datacenters — described by Anderson as “among the most complex systems that people have ever built.” Even so, Lemmon was undaunted by the subject matter. For the open-ended class project, he and his partners built a system for mapping the wiring layout for a datacenter that surprised and delighted their professor.
“As part of their project, Isaiah and his partners built a truly impressive visualization tool that showed how the different layers of switches could be organized as the size of the datacenter scaled up,” Anderson recalled. “Even though Isaiah was still just an undergrad, I thought his work was among the most ambitious and most interesting in a class of 80 students.”
“For me, it was my passion for programming and clean energy that led to where I am — finding a meaningful way to overlap those fields has been my dream,” Lemmon said. “I’m still trying to find the exact niche I want to settle in and am not yet sure if that will be in industry or academia, but my hope is some combination of the two so I can continue to do impactful research.”
The Association for Computing Machinery’s Special Interest Group on the Management of Data honored Allen School professor Dan Suciu with its 2022 Edgar F. Codd Innovations Award in recognition of his “lasting contributions to the foundations of novel data management trends.” The award recognizes a member of the ACM SIGMOD community who has made enduring and highly significant contributions to the development, understanding or use of databases and database systems over the course of their career. In the same week he was recognized for his influential body of work, Suciu also collected a Best Paper Award for his latest contribution — an indication that he has no intention of resting on his laurels.
Throughout his career, Suciu has shown a propensity for drawing deep connections between logic, database theory, algorithms and systems. According to Dan Olteanu, professor and head of the Data Systems and Theory Group at the University of Zurich, Suciu’s approach to advancing new data management paradigms has also set a new standard for research in the field.
“Dan Suciu is the most prominent researcher bridging database systems and theory,” said Olteanu. “His numerous contributions to the theory and practice of databases have fundamentally transformed a wide range of areas of research, such as semistructured data, data security, querying unreliable and inconsistent sources, probabilistic databases, data pricing, distributed and parallel query processing, and causality inference. They established clean formal foundations and practical and elegant data processing techniques. They further changed how younger generations of computer science researchers, including myself, approach research problems, and we relentlessly strive to meet the bar set by his work.”
Suciu, who holds the Microsoft Endowed Professorship in the Allen School, has been setting the bar high ever since he joined the University of Washington nearly 22 years ago. For example, in a paper appearing at the 2002 Symposium on Principles of Database Systems (PODS) Suciu and then-student Gerome Miklau (Ph.D., ‘05), now a faculty member at the University of Massachusetts Amherst, examined the complexity of containment and equivalence for a core fragment of the XPath query language for XML applications. The fragment in question covers attributes that are frequently applied in practice, specifically queries that contain branching, label wildcards and can express descendant relationships between nodes. Whereas prior work had established that efficient containment algorithms exist for any combination of two of those, Suciu and Miklau established — to their own surprise — that the problem of checking containment of all three is coNP-complete. Based on their findings, the duo designed an efficient containment algorithm capable of running in polynomial time for several cases with practical significance that involve all three. Their paper was singled out for its impact with the PODS Alberto O. Mendelzon Test of Time Awards in 2012 — one of two PODS Test of Time Awards that Suciu has received in his career so far.
In another paper that combined rigorous theoretical evaluation with real-world concerns — in this case, those of every individual who has purchased a product or service or surfed online — Suciu and his collaborators introduced a framework for the pricing of private data. This work, published in 2017, conceived of a market balancing the interests of individuals in safeguarding their personal information with the sometimes incompatible interests of companies and organizations seeking to extract value from said data for the purposes of offering more personalized services, targeting their marketing to specific interests, and direct sale to third parties. To find that elusive balance, Suciu, Miklau, and co-authors Chao Li and Daniel Yang Li drew from and expanded upon elements of differential privacy and data markets to construct a framework in which a “market maker” acts as an intermediary between individual data owners and institutional data buyers. In this scenario, the market maker responds to queries from the latter and sets prices to compensate the former based on a variety of factors, including the amount of perturbation — or noise — in the data, the option to source data from less expensive query sources, and individuals’ own risk tolerance for potential privacy loss. Suciu and the team earned a Best Paper Award at the 16th International Conference on Database Theory (ICDT 2013) for this work.
Around the same time, against the backdrop of the rapid rise in cloud computing and big data, Suciu tackled the algorithmic aspects of parallel data processing over large-scale distributed systems such as the MapReduce framework and UW’s own Myria system. Working with colleague Paul Beame of the Allen School’s Theory of Computation group and former student Paraschos Koutris (Ph.D., ‘15), now a faculty member at the University of Wisconsin-Madison, Suciu introduced a new theoretical model, Massively Parallel Computation (MPC), that separates the cost of computation from that of communication. By focusing the cost of parallel processing exclusively on the latter based on the amount of communication and the number of communication rounds, Suciu and his collaborators led a paradigm shift in how the community analyzed the complexity of distributed large-scale data queries — from run-time or the number of disk input/output operations, to the amount of data being reshuffled while maintaining server-load balance. Suciu and Koutris subsequently applied the MPC in a comprehensive survey of algorithms for different data processing tasks in collaboration with Semih Salihoglu, a faculty member at University of Waterloo.
In addition to publishing nearly 300 conference or journal papers, Suciu has contributed to multiple highly-cited books in data management. One of those, published in 2011, was an authoritative work on probabilistic databases he co-authored with Olteanu; former student Christopher Ré (Ph.D., ‘09), now a faculty member at Stanford University; and Christoph Koch, a faculty member at the École Polytechnique Fédérale de Lausanne. In the book, Suciu and his collaborators put forward novel representation formalisms and query processing techniques for modeling and processing probabilistic data used in information extraction, scientific data management, data cleaning, and other use cases that involve large volumes of uncertain data.
Three of Suciu’s aforementioned student collaborators — Miklau, Ré and Koutris — earned the ACM SIGMOD Doctoral Dissertation Award working with him. According to his UW Database Group colleague Magdalena Balazinska, Suciu’s impact in advancing new paradigms in data management is rooted not only in his vision and leadership in bridging theory and practice, but also in his devotion to developing the next generation of researchers to carry that work forward.
“As a database researcher myself, I have appreciated Dan’s approach to breaking new intellectual ground while offering a path to practical implementation,” said Balazinska, professor and director of the Allen School. “He has a knack for setting new directions for theoreticians to explore while also guiding engineers in the actual development of systems aligned with emerging trends. Not only is Dan a leader in the database research community, but he is a wonderful mentor to all who have the privilege of studying with him in addition to being a treasured colleague and friend.”
In his award talk at the SIGMOD/PODS conference in Philadelphia, Pennsylvania earlier this month, Suciu credited one of his own early mentors, Val Tannen, with “teaching me, and teaching me how to teach” and setting him on the path to his life’s work. It was Tannen who first introduced Suciu, back when he was a college student in Romania, to a new kind of mathematics that featured lattices, category theory and universal algebras — elements that Suciu found compelling and “strangely relevant” to his newfound passion for programming. Suciu later followed Tannen to the University of Pennsylvania, where he began his career in database research as a Ph.D. student working to redesign query languages grounded in mathematics — collecting the first of his many conference paper awards, at ICDT 1995, in the process.
Fast forward 27 years later, and Suciu collected his latest Best Paper Award, this time from PODS, for his work on “Convergence of Datalog over (Pre-) Semirings.” In the winning paper, Suciu and his co-authors — Mahmoud Abo Khamis and Hung Q. Ngo at RelationalAI, Reinhard Pichler at TU Wien, and Allen School Ph.D. student Remy Wang — make progress on an open problem related to enabling recursive queries beyond Boolean space as required by modern data processing and tensor computations that power applications ranging from program analysis and machine learning, to graph algorithms and linear algebra. To enable this progress, the team introduced datalogo, a powerful language for expressing recursive computations over general semi-rings.
This latest accolade brings Suciu’s career tally at influential database, data management, and related conferences to six Best Paper or Distinguished Paper Awards, three Best Demo Awards, and five Test of Time or Influential Paper Awards.
