Allen School News

Allen School professor Dieter Fox will be honored at the 32nd International Joint Conference on Artificial Intelligence (IJCAI) with the 2023 John McCarthy Award. The award is named for the eponymous scientist, widely regarded as one of the founders of the field of artificial intelligence (AI), and recognizes established researchers who have built up a distinguished track record of research excellence in AI. Fox will receive his award this week and give a presentation on his work at the conference held in Macao, S.A.R.

“Receiving the John McCarthy Award is an incredible honor, and I’m very grateful for the truly outstanding students and collaborators I had the pleasure to work with throughout my career,” Fox said. “I also see this award as a recognition of the importance the AI community places on building intelligent systems that operate in the real world.” 

Fox has made a number of key contributions to the fields of AI and robotics, developing powerful machine learning techniques for perception and reasoning, as well as pioneering Bayesian state estimation and the use of depth cameras for robotics and activity recognition. 

His research focuses on systems that can interact with their environment in an intelligent manner. Currently, most robots lack the intelligence to perceive and understand changing environments over time. They move objects in a set, programmable way. In a factory, where conditions are tightly controlled, this is a strength. Everywhere else, it’s a problem.

During his time as a Ph.D. student at the University of Bonn, Fox’s work on Markov localization tackled a fundamental problem in robotics and is now considered a watershed moment for the field. Near the start of the 21st century, researchers concentrated on the problem of tracking, giving a robot a map and its initial location. But these robots lacked true autonomy. They were unable to estimate their location and recover from mistakes out in the field — traits, importantly, displayed by a human pathfinder. 

Fox and his collaborators developed grid-based and sampling-based Bayes filters to estimate a robot’s position and orientation in a metric model of the environment. Their work produced the first approach that allowed a robot to reorient itself and recover from failure in complex and changing conditions. Fox’s pioneering work in robotics touches virtually every successful robot navigation system, be it indoors, outdoors, in the air or on streets. 

Fox’s contributions go beyond core robotics. Using a variety of data sources, including GPS, Wi-Fi signal strength, accelerometers, RFID and geospatial map information, Fox developed and evaluated hierarchical Bayesian state estimation techniques to solve human activity recognition problems from wearable sensors. With his collaborators, he demonstrated that a person’s daily transportation routines could be gleaned from a history of GPS sensor data. The work was motivated by the aim to help people with cognitive disabilities safely navigate their community without getting lost. Trained on GPS data, the wearable system assists users who get off track by helping them find public transportation to reach their intended destination. This influential work earned an Association for the Advancement of Artificial Intelligence 2004 (AAAI-04) Outstanding Paper Award, a 2012 Artificial Intelligence Journal (AIJ) Prominent Paper Award and a Ubicomp 2013 10-Year Impact Award.

In 2009, Fox began a two-year tenure as director of Intel Labs Seattle. There, he and his collaborators developed some of the very first algorithms for depth camera-based 3D mapping, object recognition and detection. Back at the University of Washington, Fox and his colleagues set an additional precedent with a separate study on fine-grained 3D modeling. Called DynamicFusion, the approach was the first to demonstrate how depth cameras could reconstruct moving scenes and objects, such as a person’s head or hands, with impressive resolution in real time. The work won a Best Paper Award from the Conference on Computer Vision and Pattern Recognition (CVPR) in 2015. 

For Fox, the McCarthy Award represents another milestone in a journey that began in his youth. As a high school student in Germany, he stumbled upon the book “Gödel, Escher, Bach: An Eternal Golden Braid” by Douglas Hofstadter. The pages, he found, flew by. When he finally closed its cover, he was spellbound. 

“From the book, I was fascinated by the ideas behind logic, formal reasoning and AI,” Fox said. “I learned that by studying computer science, I’d be able to continue to have fun investigating these ideas.”

Fox currently shares his time between the UW and NVIDIA, joining the company in 2017. He directs the UW Robotics and State Estimation Laboratory and is senior director of robotics research at NVIDIA. His work at NVIDIA stands at the cutting edge of deep learning for robot manipulation and sim-to-real transfer, bringing us ever closer to the dream of smart robots that are useful in real world settings such as factories, health care and our homes

Among his many honors, he is the recipient of the 2020 RAS Pioneer Award presented by the IEEE Robotics & Automation Society, and multiple best paper awards at AI, robotics and computer vision conferences. Fox, who joined the UW faculty in 2000, was also named a 2020 Association for Computing Machinery (ACM) Fellow, an IEEE Fellow in 2014 and a 2011 Fellow of the AAAI. Read more →

With nearly a billion unique monthly users, Wikipedia has become one of the most trusted sources of information worldwide. But while it’s considered more reliable than other internet sources, it’s not immune to bias. 

Last year, a team led by Allen School professor Yulia Tsvetkov developed a new methodology for studying bias in English Wikipedia biographies, and this spring won the 2023 Wikimedia Foundation Research Award of the Year for its efforts. The team first presented its findings at The Web Conference 2022. 

“Working with Wikipedia data is really exciting because there is such a robust community of people dedicated to improving the platform, including contributors and researchers,” Tsvetkov said. “In contrast, when you work with, for example, social media data, no one is going to go back and rewrite old Facebook posts. But Wikipedia editors revise articles all the time, and prior work has encouraged edit-a-thons and other initiatives for correcting biases on the platform.”

For the continuously evolving site, the research fills crucial content gaps in its data and how it is ultimately used. In the past, related studies focused mainly on one variable, binary gender, and lacked tools to isolate variables of interest, limiting the conclusions that could be drawn. For example, previous research involved comparing the complete sets of biographies for women and men in order to determine how gender influences their portrayals in these bios.

Tsvetkov’s team developed a matching algorithm to build more comprehensive and comparable sets, targeting not just gender but also other variables including race and non-binary gender. For instance, given a set of articles about women, the algorithm builds a comparison set about men that matches the initial set on as many attributes as possible (occupation, age, nationality, etc.), except the target one (gender).

The researchers could then compare statistics and language in those two sets of articles to conduct more controlled analyses of bias along a target dimension, such as gender or race. They also used statistical visualization methods to assess the quality of the matchings, supporting quantitative results with qualitative checks.

As a result, the researchers saw a significant difference when analyzing articles with and without their matching approach. When the approach was implemented, they found data confounds decreased — a boon for better evaluating bias in the future. 

“We did a lot of data curation to be able to include analyses of racial bias, non-binary genders, and intersected race and gender dimensions,” said lead author Anjalie Field, a professor at Johns Hopkins University who earned her Ph.D. from Carnegie Mellon University working with Tsvetkov. “While our data and analysis focus on gender and race, our method is generalizable to other dimensions.”

Future studies could further build upon the team’s methodology, targeting biases other than gender or race. The researchers also pointed to shifting the focus from the data sets to the natural language processing (NLP) models that are deployed on them. 

“As most of our team are NLP researchers, we’re also very interested in how Wikipedia is a common data source for training NLP models,” Tsvetkov said. “We can assume that any biases on Wikipedia are liable to be absorbed or even amplified by models trained on the platform.”

The study’s co-authors also included Chan Young Park, a visiting Ph.D. student from Carnegie Mellon University, and Kevin Z. Lin, an incoming professor in the University of Washington’s Department of Biostatistics. Lin earned his doctorate from Carnegie Mellon University and was a postdoc at the University of Pennsylvania when the study was published. 

Learn more about the Wikimedia Research Award of the Year here, and Tsvetkov’s research group here. Read more →

The Allen School’s Gabriel Ilharco and Ashish Sharma are among 13 students across the U.S. and England to receive 2023 J.P. Morgan AI Ph.D. Fellowships. The fellowships are part of the J.P. Morgan AI Research Awards Program, which advances artificial intelligence (AI) research to solve real-world problems.  

Ilharco, a fourth-year Ph.D. student in the Allen School’s H2Lab, is advised by professors Ali Farhadi and Hannaneh Hajishirzi. His research focuses on advancing large multimodal models as reliable foundations in AI. 

“I believe the next-generation of models will push existing boundaries through more flexible interfaces,” Ilharco said. “There is much progress to be made towards that vision, both in training algorithms and model architectures, and in understanding how to design better datasets to train the models. I hope my research will continue to help in all of these directions.”

During his fellowship, Ilharco said he hopes to continue his work in building more reliable machine learning systems. While models such as GPT-4, Flamingo and CLIP have demonstrated impressive versatility across applications, there is still room for growth. In the past decade, machine learning systems have become highly capable, particularly when performing specific tasks such as recognizing objects in images or distilling a piece of text. Yet their abilities can advance further, Ilharco said, with the end goal being a single model that can be deployed across a wider range of applications. 

To meet this challenge, Ilharco is targeting dataset design. A recent project, DataComp, acts as a benchmark for designing multimodal datasets. Ilharco was part of the research team that found smaller, more stringently filtered datasets can lead to models that generalize better than larger, noisier datasets. In their paper, the researchers discovered that the DataComp workflow led to better training sets overall. 

Ilharco and his collaborators will host a workshop centered around DataComp at the International Conference on Computer Vision 2023 (ICCV23) in October. 

“DataComp is designed to put research on datasets on rigorous empirical foundations, drawing attention to this understudied research area,” Ilharco said. “The goal is that it leads to the next generation of multimodal datasets.”

Another project introduced a framework for editing neural networks and appeared at the 11th International Conference on Learning Representations (ICLR 2023) this spring. Ilharco co-authored the paper that investigated how the behavior of a trained model could be influenced for the better using a technique called task arithmetic. In one example, the team showed how the model could produce less toxic generations when negating task vectors. Conversely, adding task vectors improved a model’s performance on multiple tasks simultaneously as well as on a single task. Ilharco and his collaborators also found that combining task vectors into task analogies boosted performance for domains or subpopulations in data-scarce environments. 

Their findings allow users to more easily manipulate a model, expediting the editing process. Because the arithmetic operations over task vectors involve only adding or subtracting model weights, they’re more efficient to compute compared to alternatives. Additionally, they result in a single model of the same size, incurring no extra inference cost. 

“We show how to control the behavior of a trained model — for example, making the model produce less toxic generations or learning a new task — by operating directly in the weight space of the model,” Ilharco said. “With this technique, editing models is simple, fast and effective.”

For Ilharco, the next wave of multimodal models is fast-approaching. He wants to be at the center of it. 

“I hope to be a part of this journey,” he said.

Sharma, also a fourth-year Ph.D. student, is advised by professor Tim Althoff in the Allen School’s Behavioral Data Science Lab. He studies how AI can support mental health and well-being. 

“I’m excited to be selected for this fellowship which will help me further my research on human-AI collaboration,” he said. “AI systems interacting with humans must accommodate human behaviors and preferences, and ensure mutual effectiveness and productivity. To this end, I am excited to pursue my efforts in making these systems more personalized.”

Sharma’s long-term goal focuses on developing AI systems that empower people in real-world tasks. His research includes work on AI to assist peer supporters to increase empathy in their communications with people seeking mental health support and exploring how AI can help users regulate negative emotions and intrusive thoughts.

Both put the user — the human being — at the center. 

“Effectively supporting humans necessitates personalization,” Sharma said. “Current AI systems tend to provide generalized support, lacking the ability to deliver experiences tailored to the specific needs of end-users. There is a need to put increased emphasis on developing AI-based interventions that provide personalized experiences to support human well-being.”

Sharma’s work with mental health experts and computer scientists was among the earliest efforts to demonstrate how AI and natural language processing-based methods could provide real-time feedback to users in making their conversations more empathetic.

At The Web Conference 2021, he and his co-authors won a Best Paper Award for their work on PARTNER, a deep reinforcement learning agent that learns to edit text to increase “the empathy quotient” in a conversation. In testing PARTNER, they found that using the agent increased empathy by 20% overall and by 39% for those struggling to engage empathetically with their conversational partners. 

“PARTNER learns to reverse-engineer empathy rewritings by initially automating the removal of empathic elements from text and subsequently reintroducing them,” Sharma said. “Also, it leverages rewards powered by a new automatic empathy measurement based on psychological theory.”

Earlier this year, Sharma was also lead author on a paper introducing HAILEY, an AI agent that facilitates increased empathy in online mental health support conversations. The agent assists peer supporters who are not trained therapists by providing timely feedback on how to express empathy more effectively in their responses to support seekers in a text-based chat. HAILEY built upon Sharma’s work with PARTNER. 

In addition, Sharma and his collaborators recently won an Outstanding Paper Award at the 61st annual meeting of the Association for Computational Linguistics (ACL 2023) for developing a framework for incorporating cognitive reframing, a tested psychological technique, into language models to prompt users toward healthier thought processes. With cognitive reframing, a person can take a negative thought or emotion and see it through a different, more balanced perspective. 

With a focus on people and process, Sharma sees how his research area can continue to grow. He said he hopes to advance AI’s ability to personalize to the user, while also remaining safe and secure. 

“Utilizing my experience in designing and evaluating human-centered AI systems for well-being, I will investigate how such systems can learn from and adapt to people’s contexts over time,” Sharma said. “I’ve always been fascinated by technological efforts that support our lives and well-being.” Read more →

Allen School researchers took home multiple Best Paper and Outstanding Paper Awards from the 61st Annual Meeting of the Association for Computational Linguistics (ACL) held in Toronto last month. Their research spanned a number of projects aimed at enhancing the performance and impact of natural language models, including how artificial intelligence (AI) processes humor, the impact of built-in political biases on model performance, AI-assisted cognitive reframing to support mental health, identifying “WEIRD” design biases in datasets and how to imbue language models with theory of mind capabilities. Read more about their contributions below.

Best Paper Awards

Do Androids Laugh at Electric Sheep? Humor ‘Understanding’ Benchmarks from The New Yorker Caption Contest

Allen School professor Yejin Choi and her collaborators earned a Best Paper Award for their study exploring how well AI models understand humor, challenging these models with three tasks involving The New Yorker Cartoon Caption Contest.

The tasks included matching jokes to cartoons, identifying a winning caption and explaining why an image-caption combination was funny. For an AI model, it’s no joke. Humor, the authors point out, contains “playful allusions” to human experience and culture. Its inherent subjectivity makes it difficult to generalize, let alone explain altogether. 

“Our study revealed a gap still exists between AI and humans in ‘understanding’ humor,” said Choi, who holds the Wissner-Slivka Chair at the Allen School and is also senior research manager for the Allen Institute for AI’s MOSAIC project. “In each task, the models’ explanations lagged behind those written by people.” 

The team applied both multimodal and language-only models to the caption data. Compared to human performance, the best multimodal models scored 30 accuracy points worse on the matching task. 

Even the strongest explanation model, GPT-4, fell behind. In more than two-thirds of cases, human-authored explanations were preferred head-to-head over the best machine-authored counterparts. 

Future studies could focus on other publications or sources. The New Yorker Cartoon Caption Contest represents only a “narrow slice” of humor, the authors note, one that caters to a specific audience. New research could also explore generating humorous captions by operationalizing feedback produced by the team’s matching and ranking models. 

The study’s authors also included Jack Hessel and Jena D. Hwang of AI2, professor Ana Marasović of the University of Utah, professor Lillian Lee of Cornell University, Allen School alumni Jeff Da (B.S., ‘20) of Amazon and Rowan Zellers (Ph.D., ‘22) of OpenAI and Robert Mankoff of Air Mail and Cartoon Collections. 

From Pretraining Data to Language Models to Downstream Tasks: Tracking the Trails of Political Biases Leading to Unfair NLP Models

Allen School professor Yulia Tsvetkov, Ph.D. student Shangbin Feng and their collaborators earned a Best Paper Award for their work focused on evaluating pretrained natural language processing (NLP) models for their political leanings and using their findings to help combat biases in these tools. 

To do this, they developed a framework based on political science literature for measuring bias found in pretrained models. Then they analyzed how these biases affected the models’ performance in downstream social-oriented tasks, such as measuring their ability to recognize hate speech and misinformation. 

They found that bias and language are difficult to separate. Both non-toxic and non-malicious data, they note, can cause biases and unfairness in NLP tasks. If political opinions are filtered from training data, however, then questions arise concerning censorship and exclusion from political participation. Neither is an ideal scenario.

“Ultimately, this means that no language model can be entirely free from social biases,” Tsvetkov said. “Our study underscores the need to find new technical and policy approaches to deal with model unfairness.”

The consistency of the results surprised the team. Using data from Reddit and several news sources, the researchers found that left-leaning and right-leaning models acted according to form. The left-leaning models were better at detecting hate speech towards minority groups, while worse at detecting hate speech towards majority groups. The pattern was reversed for right-leaning models. 

In evaluating data from two time periods — before and after the 2016 U.S. presidential election — they also discovered a stark difference between the levels of political polarization and its attendant effect on the models’ behavior. With more polarization comes more bias in language models. 

Future studies could focus on getting an even more fine-grained picture of the effect of political bias on NLP models. For example, the authors note that being liberal on one issue does not preclude being conservative on another. 

“There’s no fairness without awareness,” Tsvetkov said. “In order to develop ethical and equitable technologies, we need to take into account the full complexity of language, including understanding people’s intents and presuppositions.”

The study’s co-authors also included Chan Young Park, a visiting Ph.D. student from Carnegie Mellon University, and Yuhan Liu, an undergraduate at Xi’an Jiaotong University.  

Outstanding Paper Awards

Cognitive Reframing of Negative Thoughts through Human-Language Model Interaction

Ph.D. students Ashish Sharma and Inna Wanyin Lin and professor Tim Althoff, director of the Allen School’s Behavioral Data Science Group, were part of a team that won an Outstanding Paper Award for their project investigating how language models can help people reframe negative thoughts and what linguistic attributes make this process effective and accessible. 

Working with experts at Mental Health America, the team developed a model that generates reframed thoughts to support the user. For example, the model could produce reframes that were specific, empathic or actionable — all ingredients for a “high-quality reframe.” The study was the first to demonstrate that these all make for better reframes, Althoff said, and the team illustrated this with gold standard randomized experiments and at scale. 

The research has already seen real-world impact. Since its introduction late last year, the team’s reframing tool has had more than 60,000 users. 

“The findings from this study were able to inform psychological theory — what makes a reframe particularly effective?” Sharma said. “Engaging with real users helped us assess what types of reframes people prefer and what types of reframes are considered relatable, helpful and memorable.”

Those “high-quality reframes” could be particularly helpful for those who lack access to traditional therapy. The team pointed out several obstacles to care, including clinician shortages, lack of insurance coverage and stigmas surrounding mental health, that served as motivations for the study. 

The model can also be integrated into existing therapy workflows, Sharma added, helping both clients and therapists in the process. For example, therapists often assign “homework” to clients, asking them to practice cognitive reframing, a technique by which a person can picture a negative thought through a different, more balanced perspective. 

But many clients report having difficulty in applying those techniques following a session. Sharma and Althoff said the team’s reframing tool can provide support in those moments. 

“It turns out that often our thoughts are so deep-rooted, automatic and emotionally triggering that it can be difficult to reframe thoughts on our own,” Althoff said. “This kind of research not only helps us improve our intervention itself, but could also inform how clinicians teach these skills to their clients.”

The study’s co-authors also included Kevin Rushton, Khendra G. Lucas and Theresa Nguyen of Mental Health America, Allen School alum David Wadden (Ph.D., ‘23) of AI2 and Stanford University professor and clinical psychologist Adam S. Miner.

Minding Language Models’ (Lack of) Theory of Mind: A Plug-and-Play Multi-Character Belief Tracker

Choi and Tsvetkov worked with Allen School Ph.D. students Melanie Sclar and Peter West and collaborators on SymbolicToM, an algorithm that improves large language models’ abilities to reason about the mental states of other people that earned the team an Outstanding Paper Award. The team also won the Outstanding Paper Award at the ToM Workshop at the 2023 International Conference on Machine Learning (ICML) for this work. 

Theory of mind (ToM), or the ability to reason about others’ thoughts and intentions, is a key part of human intelligence. But today’s AI models lack ToM capabilities out of the box. Prior efforts at integrating ToM into language models required training, with existing reading comprehension datasets used for ToM reasoning remaining too simplistic and lacking diversity. 

“This implied that models trained solely on this data would not perform ToM reasoning,” Sclar said, “and rather only mimic these skills for the simplistic data they were trained on.”

Enter SymbolicToM. Without requiring ToM-specific training, the decoding-time algorithm takes a divide-and-conquer approach. SymbolicToM splits a given problem into subtasks, Sclar said, solving each with off-the-shelf large language models. The result is a better, more robust model. 

“We knew from the get-go that our approach needed to focus on having good generalization capabilities, and thus would benefit from not requiring training,” Sclar said. “SymbolicToM is to the best of our knowledge the first method for theory of mind reasoning in natural language processing that does not require any specific training whatsoever.”

The team tasked SymbolicToM with answering reading comprehension questions based on a story featuring multiple characters. They tracked each character’s beliefs, their estimation of others’ beliefs and higher-order levels of reasoning through graphical representations. In doing so, the models could reason with more precision and interpretability.  

“Our method in particular is not focused on training neural language models, but quite the opposite: given that we have imperfect language models trained with other objectives in mind, how can we leverage them to dramatically improve theory of mind performance?” Tsvetkov said. “This is key because data with explicit theory of mind interactions are scarce, and thus training directly is not a viable option.”

Sclar pointed to potential avenues for future applications, including education and business. For example, AI agents with ToM reasoning skills could assist in tutoring applications, providing a deeper understanding of students’ knowledge gaps and designing tests based on their mental model of each student. 

Another instance involves negotiation strategy. If AI agents can intuit what each party hopes to achieve and how much they value certain aspects of a deal, Sclar said, they can provide support in reaching a fair consensus. 

“Imbuing neural language models with ToM capabilities would improve these models’ potential on a wide range of applications,” Sclar said, “as well their understanding of human interactions.”

The study’s authors also included visiting Ph.D. student Sachin Kumar of Carnegie Mellon University and professor Alane Suhr of the University of California, Berkeley. 

NLPositionality: Characterizing Design Biases of Datasets and Models

Ph.D. student Sebastin Santy, professor Katharina Reinecke and their collaborators won an Outstanding Paper Award for devising a new framework for measuring design biases and positionality in NLP datasets that provides a deeper understanding of the nuances of language, stories and the people telling them. 

“Language is a social phenomenon,” Santy said. “Many in the NLP field have noticed how certain datasets and models don’t work for different populations, so we felt it was the right time to conduct this large-scale study given these gaps and with the right kind of platform.” 

That platform, LabintheWild, provides more reliable data from a more diverse set of users. Reinecke, one of the platform’s co-founders and director of the Wildlab at the Allen School, noted that as opposed to Mechanical Turk, a popular paid crowdsourcing site, LabintheWild collects results from a greater pool of countries. 

With LabintheWild, the personal is emphasized over the pecuniary. After completing a study, users can see personalized feedback and compare their results with others’ performance on the platform.

This feedback is eminently shareable, Reinecke added, increasing its reach. The researchers’ recent study collected 16,299 annotations from 87 countries — one of the first NLP studies to reach that scale. They applied their framework, called NLPositionality, to LabintheWild’s vast participant pool, implementing users’ annotations from existing datasets and models for two tasks: social acceptability and hate speech detection. 

Their findings aligned with Reinecke’s previous work, which shows that technology is often designed for people who are Western, Educated, Industrialized, Rich and Democratic, or “WEIRD.” 

“WEIRD bias is well-known in psychology and our hypothesis was that we might find similar results in AI as well, given most of the recent advances make use of mostly English data from the internet and filter for ‘high-quality,’ ” said Reinecke, who holds the Paul G. Allen Career Development Professorship. “While we had a feeling that there would be Western bias because of how most of the datasets are curated in the Western Hemisphere, we did not expect it to be this pronounced.”

The study’s co-authors also included Allen School alumni Maarten Sap (Ph.D., ‘21) and Jenny Liang (B.S., ‘21), now professor and Ph.D. student, respectively, at Carnegie Mellon University, and Ronan Le Bras of AI2.  Read more →

In the days before single-cell RNA sequencing, researchers investigating the mechanisms and treatment of disease had to make do with running experiments on bulk cell profiles created by taking tissue samples and grinding them up, “sort of like putting them in a blender,” in the words of Allen School Ph.D. student Ethan Weinberger.

That milkshake may have brought all the biomedical scientists to the lab, but the bulk sequencing technique limited them to studying aggregations of populations of cells, with no way to distinguish among individual cell types. Nowadays, researchers can take measurements at the level of individual cells, enabling the exploration of such finer-grained distinctions and advancing our understanding of various biological functions. But without the right computational tools, even single-cell datasets can yield distinctions without a difference.

Weinberger is a member of the Allen School’s AIMS Lab, where he works with fellow Ph.D. student Chris Lin and professor Su-In Lee to leverage advances in artificial intelligence to help scientists get the most out of these increasingly robust datasets. In a paper published this week in Nature Methods, the team introduced ContrastiveVI, a deep learning framework for applying a powerful technique called contrastive analysis, or CA, to single-cell datasets to disentangle variations in the target, or treatment, cells from those shared between target and control cells when running experiments. 

“Scientists want to investigate questions like ‘How does perturbing this particular gene affect its response to a pathogen?’ or ‘What happens when I hit a diseased cell with such-and-such a drug?’,“ explained Weinberger. “To do that, they need to be able to isolate the variations in the cell data caused by that perturbation or that drug from those that are shared with a control dataset. But existing models can’t separate those out, which might lead someone to draw erroneous conclusions from the data. ContrastiveVI solves that problem.”

CA has proven effective at this type of isolation in other contexts, but its utility in relation to single-cell datasets has so far been limited. That’s because existing computational models for analyzing single-cell data mostly rely on a single set of latent variables to model all variations in the data, effectively lumping them all together and precluding the ability to perform CA.

ContrastiveVI is the first deep learning model designed for performing CA on single-cell data. Unlike other approaches, the ContrastiveVI model explicitly separates latent variables into two categories, each with their own encoding function: shared variables, or those that are found in both the target and control cells, and salient variables, which are found exclusively among the target cells. 

It is that second category that will excite scientists testing potential cancer drugs or analyzing the role of gene expression in the body’s response to disease. 

“ContrastiveVI effectively distinguishes the factors that are salient — that is, relevant — to an experiment from confounding factors. This enables us to capture variations that are unique to the treated cells,” said Lee, senior author of the paper and holder of the Paul G. Allen Career Development Professorship in the Allen School. “ContrastiveVI will reveal tiny but important variations in the data that may be obscured by other models.”

Lee and her co-authors validated ContrastiveVI using real-world datasets with previously verified results as their ground truth. In one experiment, the researchers applied ContrastiveVI to a target dataset of measurements taken from two dozen cancer cell lines treated with idasanutlin. This small-molecule compound has shown therapeutic potential owing to its activation of a tumor-suppressing protein in wild type — that is, unmutated — TP53 genes. The team used ContrastiveVI to analyze data on both wild type and mutated TP53 cell lines, which are non-responsive to idasanutlin, using a background dataset from the same cell lines treated with a different compound, dimethyl sulfoxide, as the control. 

“A good result — one that agreed with prior knowledge — would show separation by cell line accompanied by increased mixing of treatment and control cells in the shared latent space, but mixing across mutant cell lines with clear separation based on mutation status in the salient latent space,” said Lin, co-lead author of the paper with Weinberger. “And that is exactly what we observed. In addition, our model indicated a separation between wild-type cell lines in the salient space that suggested a differential response to treatment, which spurred us to run additional analyses to identify the specific genes that contribute to those variations.”

Such findings, which could build upon prior knowledge and lead scientists to new hypotheses, is precisely the sort of progress Lin and his colleagues hope their model will support. In another demonstration of ContrastiveVI’s potential, the researchers applied the model to a dataset drawn from intestinal epithelial cells of mice displaying variations in gene expression due to infection with the bacteria Salmonella or the parasite H. polygyrus (H. poly), a type of roundworm, using healthy cells as the control. Once again, the model aligned with expectations by separating along cell type and mixing across infections in the shared latent space, while largely mixing across cell types and separating by pathogen in the salient latent space.

Like the cancer cell example, the pathogen infection experiment also yielded unexpected patterns that prompted the team to analyze further. These patterns included differences in the upregulation of multiple genes between H. poly–infected tuft cells and other infected cell types that may have been masked in prior experiments — and could point to a distinctive role in the body’s immune response.

The researchers also explored how the model could be adapted to isolate variations in multimodal single-cell datasets, such as a combination of RNA and surface protein expression data in CRISPR-perturbed cells. They layered their CA modeling techniques onto TotalVI, a deep generative model developed to analyze joint RNA-protein datasets, to create TotalContrastiveVI. In a series of experiments, they showed how their extended model could be used to identify clusters of cells in the salient latent space and apply downstream analysis to identify patterns that warranted further investigation.

TotalContrastiveVI may be a proof of concept, but the underlying model is no mere demonstration project. The team designed ContrastiveVI to make it easy for researchers to integrate the tool into existing workflows.

“Our software is essentially plug and play,” noted Lin. “Computational biologists can deploy ContrastiveVI right now in conjunction with standard tools in the field such as Scanpy to begin exploring single-cell datasets in greater detail than they could before.”

Those details could yield new hypotheses that could, in turn, lead to new biomedical breakthroughs.

“There are so many contexts in which scientists would want to do this,” said Weinberger. “People were already excited by the potential of single-cell datasets. With ContrastiveVI, they can unlock even more insights and expand our knowledge of the mechanisms and treatment of disease.

“To borrow a popular metaphor in biomedical circles: before, we had a smoothie; now we can zoom in on each part of the corresponding fruit salad.”

Read the paper in Nature Methods here. Read more →

A chance encounter helped Paul Beame, Paris Koutris (Ph.D., ‘15) and Dan Suciu create a model that aids scientists in understanding some of the deeper nuances surrounding big data management. Beame was passing by Suciu’s office one day as the latter and his advisee Koutris, a Ph.D. student in the Allen School at the time, were discussing research on parallel query processing, a shared idée fixe among the trio. 

They were thinking about how to model the complexity of new modes of computation used in practice, such as MapReduce, a creation of Allen School alum Jeff Dean (Ph.D., ‘96) and Sanjay Ghemawat that was developed to run major systems at Google. MapReduce, however, involved a specific set of parallel primitives that limited what types of operations could be captured. 

For Koutris and Suciu, a professor in the University of Washington’s Database Group, more was possible. They wanted a theoretical model, Suciu recalled, that could also capture other computations that could be performed more efficiently with other primitives. They also wanted to better understand “the fundamental limits” of these systems’ computing capabilities. They just needed an added spark, another stroke of inspiration. 

Along came Beame, a professor in the Allen School’s Theory of Computation group, and the rest was history. 

“It turned out that Paul had already been mulling over the general question of modeling MapReduce computations but there had been an aspect missing in his thinking that was a feature in my and Paris’ work on parallel database query processing,” Suciu said. “Even in the discussion that very first day, it became clear that the fundamental limitation that the model needed to capture was the communication of data that processors require in order to solve computational problems.” 

Beame added that the meeting was made possible by the collaboration and openness of the culture in CSE. Suciu’s door was open; his and Koutris’ minds, even more so. 

“We value collaboration and the mixing of ideas among different research areas enough that we put faculty from different areas near each other, rather than cloistering faculty in different zones by research area,” Beame said. “Dan’s office is on the way to mine just three doors away.”

Their work has helped crystallize what’s possible in the world of big data processing. Last month, the trio’s paper “Communication Steps for Parallel Query Processing” earned the 2023 ACM PODS Alberto O. Mendelzon Test-of-Time Award for providing a deeper understanding of the complexity behind database query evaluation. The award was announced at the 2023 ACM SIGMOD/PODS conference, jointly organized by the Association for Computing Machinery’s Special Interest Group on Management of Data and the Symposium on Principles of Database Systems in Seattle last month. 

Koutris joined the University of Wisconsin-Madison as a professor of computer science after graduating from the Allen School 2015.

The three continued to work together over a series of subsequent papers, investigating different questions related to their initial research. One involved data skew. By allowing additional communication rounds, the group theorized, they would be able to compute more complex queries. 

“That, indeed, turned out to be true,” Suciu said. “However, to our surprise, additional communication rounds also allowed the same query to process data that is skewed.” 

For example, Suciu added, there may be many more records referring to former president Barack Obama than records referring to a less famous individual. These “heavy hitters” pose a major problem for distributed computation, he continued, because they tend to overload a server and slow down the entire computation. 

In their Test of Time paper, the trio showed the heavy hitters’ impact could be blunted. By using additional rounds of communication, these records’ effects could be mitigated when distributed across multiple servers. 

“To our surprise,” Suciu said, “the additional communication rounds turned out to be very effective in decreasing the total cost of computation on skewed data.”

For the team, their model, dubbed Massively Parallel Communication (MPC), is yet another marker in the race toward measuring the complexity of query processing systems. During the early 2000s, the database and data processing environment underwent sweeping changes, with several systems, including MapReduce, popping up that could process vast swaths of data by using thousands of commodity systems. The process was unprecedented — and relatively cheap. 

“At around the same time Amazon provided the first cloud services, allowing users to access massively distributed systems,” Suciu said. “The traditional way to measure the complexity of a query was in terms of the number of accesses to the disk. However, these new, massively distributed systems were using a sufficiently large number of servers to store the entire data in main memory.” 

With progress came a new set of challenges. Now, the source of complexity, Suciu said, was the communication cost among the servers. It required a new theoretical approach to understand the contours of a changing landscape. 

“Once we had the model,” Beame said, “our focus was on understanding the complexity and best ways to solve natural and important classes of problems in database query evaluation.”

The team started with a simple model consisting of a large number of servers, with each featuring arbitrary computational power. The only limitation was the amount of data exchanged among the servers and the number of communication rounds.  

“Our breakthrough came when we were able to explain in this model why some queries were easier to compute than others,” Suciu said, “and to precisely characterize the communication cost of the queries.”

Their breakthrough has far-reaching implications. Thanks to the team’s research, systems developers and data analysts can better understand the limitation of processing complex queries on massively distributed systems. While it’s not the first model to attempt to capture synchronous distributed parallel computation, its elegance and simplicity have made it popular and able to withstand the test of time. 

“The MPC model was very clean and natural to understand and think about,” Beame said. “The introduction of the model was quite timely and the MPC terminology has become widely adopted and used.”  Read more →

Take a generous helping of mathematical brilliance, cover it in copious amounts of curiosity about the most vexing problems underpinning computer science, add a generous dash of humility, and what do you get? 

Shayan Oveis Gharan, a professor in the Allen School’s Theory of Computation group, combines all the essential ingredients of a trailblazing researcher who, as his colleagues will attest, also happens to be a genuinely nice guy. The combination has also proved to be a genuine recipe for success, as he has racked up a series of accolades in theoretical computer science since he arrived at the University of Washington in 2015. His most recent honor, the Stephen Smale Prize from the Society for the Foundations of Computational Mathematics (FoCM), celebrated Oveis Gharan’s “breakthrough results on the applications of algebraic and spectral methods to the design of algorithms and to combinatorial optimization” that have made him “the architect of surprising and profound discoveries on foundational problems in computing.”

Oveis Gharan began acquiring the building blocks of those career triumphs as a middle school student in Iran, when he encountered the book “Mathematics of Choice: How to Count without Counting” by the Canadian-American mathematician Ivan Niven. Under Niven’s written tutelage, the young Oveis Gharan learned how to quickly count combinatorial objects — say, the number of ways one could assemble a basketball team consisting of 10 players and a captain from a class of 30 students — on paper. The lesson was a slam dunk, and counting problems still comprise a major portion of his research years later — only now, he’s designing computer algorithms to handle more sophisticated problems than the possible permutations with himself at point guard. 

Later, as he prepared to do battle in regional and world informatics competitions, a teenaged Oveis Gharan practiced with hundreds of combinatorial and graph theoretic problems drawn from past Russian Mathematics Olympiads. That practice paid off, as he took home a silver medal from the 2003 Central European Olympiad; a year later, he won gold at the International Olympiad.

“Those experiences are the backbone of most problem-solving techniques I still use today,” Oveis Gharan noted, pointing out that old approaches can come in handy when it comes to new problems. It’s one of the aspects he loves most about his chosen line of work.

“One of the amazing characteristics of research in discrete mathematics and combinatorics is that there is rarely a unified theme to approach hard problems,” he explained. “One often needs to cook up a new method from scratch. So in one sense, it feels like we are going to fight with a challenging problem empty-handed, but in reality, previous work on related problems can offer a menu of ideas with which to approach this other problem.”

One challenge that Oveis Gharan found he could really sink his teeth into is the infamous Traveling Salesperson Problem, which he first encountered over a decade ago as a Ph.D. student at Stanford University. There, he and his colleagues set out to design an improved approximation algorithm for metric TSP. 

At first, the team thought they had the solution — until they realized they didn’t.

“About halfway through we figured some of our intermediate conjectures were wrong,” Oveis Gharan recalled. “So, we made the problem simpler and instead only managed to prove that the algorithm breaks the long-standing barrier for a special family of metrics called ‘graph metrics.’ It wasn’t until two years ago that I and another group of co-authors finally achieved a result for all metrics.”

The aforementioned result was the first performance improvement in metric TSP in more than 40 years. Along the way, Oveis Gharan contributed to what co-author and Allen School professor Anna Karlin has described as a “deep mathematical machinery” mixing elements of graph theory and probability theory that researchers have since applied to other open problems. Among the tools in this expanded mathematical toolbox were the use of maximum entropy sampling, new theorems related to the geometry of polynomials, Strong Rayleigh probability distributions and negative dependence, and new insights into the combinatorial structure of the minimum and near-minimum cuts of a graph. Oveis Gharan and his co-authors, including Ph.D. student Nathan Klein, subsequently used the latter to build on their initial result by showing the integrality gap of the subtour linear programming relaxation for TSP is below 3/2 last year.

What is it about TSP that he finds so alluring? As Oveis Gharan explains it, TSP is different from most computational problems encountered by theoreticians, like the graph coloring problem, that require one to satisfy a range of local constraints. With TSP, one has to satisfy both a set of local constraints and a global constraint — connectivity — simultaneously.

“Oddly enough, each of these two sets of constraints is easy to satisfy optimally on their own, but the challenge is to satisfy both,” he said. “The quest in studying TSP is that you want to construct solutions which are locally correct while globally connected.”

For Oveis Gharan, satisfying that dual challenge is where the rubber meets the road.

”Think of a driver going from Seattle to San Francisco. They need to keep an eye on the road to make sure they are ‘locally’ driving safely — i.e., not ramming into the next car and not driving out of bounds,” he said. “But they also need to keep the bigger picture of the route in mind, choosing the right highway at every junction. Now in this example, perhaps, the bigger picture is easy to keep in mind when there’s only a single highway, I-5, running all the way south. But imagine how difficult it would be with millions of possible roads to choose from, and no GPS! That is similar to the dilemma in designing an algorithm for TSP.”

Despite his devotion to that problem, Oveis Gharan is also driven to tackle other challenges. For example, he is widely known for his work analyzing the Markov Chain Monte Carlo (MCMC) technique for sampling from high dimensional distributions and as a method for studying large, complicated sets. As part of that work, Oveis Gharan and his collaborators — including former student Kuikui Liu (Ph.D., ‘23), who will join the faculty of MIT this fall — developed the theory of spectral independence, a revolutionary approach for approximate sampling of Markov chains that has implications for computational biology, machine learning, physics and more. 

Oveis Gharan and his colleagues leveraged that approach to produce the first efficient approximation algorithm for counting the bases of a matroid and simultaneously proved a 30-year-old conjecture concerning the minimum edge expansion of the bases exchange graph of any matroid. In another paper of the same series, he and his co-authors answered another open question that had gone unanswered for nearly three decades by proving that an algorithmic tool used in statistical physics called Glauber dynamics mixes in polynomial time to generate a random independent set for any graph up to the phase-transition threshold.

The aforementioned work relates to expander graphs — which, in Oveis Gharan’s view, are “one of the most extraordinary inventions in mathematics.” He is keen to further explore the theory of high-dimensional expanders, which have numerous practical applications in computing.

“On one hand, these graphs are as sparse as, say, a cycle; on the other hand, they preserve almost all properties of a complete network,” Oveis Gharan explained. “If one wants to build a sparse routing network that will be as reliable as possible against node or connection failures, the best design is to use an expander graph. High dimensional expanders are a generalization of expander graphs to hypergraphs, which have been a subject of intense study over the last decade leading to several breakthroughs, from improved analysis of MCMC algorithms, to the construction of new locally testable codes.”

To someone whose work is typically celebrated in optimization and algorithm design circles, the Smale Prize came as a pleasant surprise. 

“Although not the community in which I normally publish my research, I am truly honored and amazed that my work has been recognized by leaders in computational mathematics,” he said. “This certainly motivates me and my research group to pursue a deeper understanding of problems at the intersection of math, computer science and efficient computation.”

The Smale Prize — named for Stephen Smale, one of the founding members of FoCM — is awarded every three years. The organization formally honored Oveis Gharan, who joins rarefied company as only the fifth recipient since the prize’s inception, during its annual conference in Paris last month. Read more →

There is an old saying that perception is everything, and with regard to human senses and computer models that attempt to demonstrate how human sensory systems work, everything is highly complex. This includes our system of visual perception, which allows humans to interact with the dynamic world in real time. Because of biological constraints and neural processing delays, our brains must “fill in the blanks” to generate a complete picture. 

Allen School professor Rajesh Rao has dedicated much of his career to the discovery of computational principles that underlie the human brain’s abilities to learn, process and store information. Recently, Rao and Allen School third-year Ph.D. student Linxing Preston Jiang have focused on understanding how our brains fill in those blanks by developing a computational model to simulate how humans process visual information. Their paper presenting their findings, “Dynamic Predictive Coding Explains Both Prediction and Postdiction in Visual Motion Perception,” will receive a Best Paper Award in the Perception and Action category at the Cognitive Science Society conference (COGSCI 2023) later this month. 

“At a high level, our work supports the idea that what we consciously perceive is an edited timeline of events and not a single instant in time. This is because the brain represents events at multiple timescales,” explained Rao, the Cherng Jia and Elizabeth Yun Hwang Professor in the Allen School and UW Department of Electrical & Computer Engineering and co-director of the Center for Neurotechnology. “At the highest levels of the brain’s networks, the neural representations cover sensations in the past, present and predicted future. This explains some seemingly strange perceptual phenomena like postdiction and the flash lag illusion, where what happens in the future can affect how you perceive what is happening now.”

In their paper, Rao and Jiang hypothesized that human sensory systems encode entire sequences, or timelines, rather than just single points in time to aid perception. To test this notion, they trained a neural network model, dynamic predictive coding (DPC), to predict moving image sequences. Since DPC learns hierarchical representations of sequences, the duo found the system is able to predict the expected perceptual trajectory while compensating for transmission delays. When events deviate from the model’s expectation, the system retroactively updates its sequence representation to catch up with new observations.

In the two-level DPC model, lower-level neural states predict the next sensory input as well as the next state, while higher-level neurons predict the transition dynamics between those lower-level neural states. This enables higher-level neurons to predict entire sequences of lower-level states. This approach is grounded in predictive coding, a theory in neuroscience that suggests the brain has separate populations of neurons to encode the best prediction of what is being perceived and to identify errors in those predictions.

“Our model fits within the class of predictive coding models, which are receiving increasing attention in neuroscience as a framework for understanding how the brain works,” Rao noted.  “Generative AI models like ChatGPT and GPT-4 that are trained to predict the next word in a sequence are another example of predictive coding.”

The team observed that when DPC’s visual perception relies on temporally abstracted representations — what the model has learned through minimizing prediction errors — many known predictive and postdictive phenomena in visual processing emerged from the model’s simulations. These findings support the idea that visual perception relies on the encoding of timelines rather than single points in time, which could influence the future direction of neuroscience research. 

“While previous models of predictive coding mainly focused on how the visual system predicts spatially, our DPC model focuses on how temporal predictions across multiple time scales could be implemented in the visual system,” explained Jiang, who works with Rao in the Allen School’s Neural Systems Lab. “If more thoroughly validated experimentally, DPC could be used to develop solutions such as brain-computer interfaces, or BCIs, that serve as visual prostheses for the blind.”

This is just the beginning for research on DPC. A future avenue of investigation might include assessing more than the two levels of temporal hierarchies Jiang and Rao examined in the current paper with videos that have more complicated dynamics. This could facilitate comparisons of responses in DPC models with neural recordings to further explore the neural basis of multi-scale future predictions. 

“Predicting the future often involves taking an animals’ own actions into account, which is a significant aspect of learning that DPC does not address,” explained Jiang. “Augmenting the DPC model with actions and using it to direct reward-based learning would be an exciting direction that has connections to reinforcement learning and motor control.” 

COGSCI brings together scholars from around the world to understand the nature of the human mind. This year’s COGSCI conference will take place in Sydney, Australia. Read more →

The possibility of a greener future inspired Meghna Shankar, a recent Allen School alum, to study alternative energy and help wean society off of fossil fuels. Her interdisciplinary academic career, she said, gave her the perspective to see the bigger picture, while also giving her the tools to make her goals a reality. 

“Since the world of scientific research is relying more and more on computing,” she said, “the skills I learn in my CS classes have come in handy in a variety of situations.”

For Shankar, who graduated in June with bachelor’s degrees in computer science and comprehensive physics, the discipline has been another path to a deeper understanding of the physical world. She was recently named a 2023 Dean’s Medalist in the Natural Sciences by the College of Arts & Sciences in recognition of her many scholastic achievements.

This fall, she will begin a doctoral program in experimental condensed matter physics at MIT, where she hopes to research topics that could have clean energy applications, such as batteries, fuel cells and solar panels. 

“UW was one of the best places for me to gain undergraduate research experience,” Shankar said. “I had the opportunity to work with amazing graduate students and postdocs, which helped me learn about the research process and the possibility of pursuing a Ph.D.”

Shankar’s Dean’s Medal selection comes on the heels of a prolific undergraduate career. As a first-year student, she began research at professor Cody Schlenker’s experimental chemistry laboratory, investigating how to increase the efficiency of solar cells. That experience led to an internship at the Pacific Northwest National Laboratory, where her work resulted in her receiving the prestigious Mary Gates Research Scholarship. In her junior year she started working in the experimental condensed matter physics lab of professor Xiaodong Xu. She has also served as a teaching assistant in introductory physics courses in addition to the Allen School’s first year seminar.

Beyond the classroom, Shankar has demonstrated a commitment to both the arts and her community. The talented violist joined the UW Symphony Orchestra as a freshman and has performed with the ensemble since then. 

“Orchestra has been an invaluable part of my UW experience,” she said. “In my freshman year, some of my first friends at UW were in the viola section.”

Shankar also helped lead Women+ in Physics, a student organization that promotes inclusivity in the sciences. Senior students had previously begun the process to incorporate the group into the physics department, but due to the pandemic, they faced challenges in getting off the ground. Shankar and her friends revitalized the organization and have since hosted many community and professional development events.  

“I have really enjoyed this experience, as it has brought me closer to my friends and made me feel like I am making an impact in the physics community,” she said. “The physics department has been incredibly supportive of us as well.”

Shankar credited her friends and family for encouraging her and fostering an environment for success. She will keep them and her UW experience in mind, she said, when she heads to MIT later this year. 

“Without the support of and collaboration with my friends and classmates, I would not have felt the same motivation to go to class and study every day,” she said. “I also want to recognize my mother and my father. They have been a huge inspiration to me to work through adversity and be committed to my goals.”

Read more about Shankar’s achievements here. Read more →

For faculty members who are at the start of their research journey, the National Science Foundation’s CAREER Awards are one of the most prestigious honors recognizing early-career scholarship and supporting future leaders in their respective fields. The latest Allen School recipients are no exception. From using machine learning to fight implicit bias to devising new architectures that bridge electronics and biology, here are eight rising stars who are advancing the field of computing at the University of Washington and reaching new heights.

Tim Althoff: Advancing behavioral data science to improve health and well-being

About a fifth of the world’s children and adolescents experience a mental health condition, according to the World Health Organization, with depression and anxiety costing the global economy $1 trillion each year. Between 2007 and 2017, Pew Research Center found that the total number of teenagers who reported experiencing depression increased by 59%. 

In “Realizing the potential of behavioral data science for population health,” Tim Althoff, director of the Behavioral Data Science Lab, aims to address rising mental health challenges by developing computational tools and utilizing behavioral health data. Though health-related data from phones, watches, fitness trackers and apps have become more widely available, integrating and modeling that data have remained difficult. Althoff’s project will attempt to develop a unified representation learning framework that effectively generalizes to new users, while also being highly predictive, robust and private without revealing private identifying information. 

His team will evaluate the framework’s performance across a number of health applications, including behavioral monitoring of influenza and COVID-19 symptoms, as well as personalizing sleep and mental health interventions. One project, for instance, uses AI to enable more empathic conversations among peer-to-peer support networks.

With this data in hand, Althoff’s group is seeking to expedite the development of new behavioral health research and related applications. They are also focused on helping scientists and health professionals learn more about the impact of behavioral health conditions. 

“Despite the significant potential of increasingly available data, broad and tangible impacts have yet to be realized, in part due to the unique challenges of integrating and modeling a broad range of behavioral and health data,” Althoff said. “This project seeks to address these challenges by developing and sharing computational tools that will enable researchers, clinicians and practitioners to improve mental health care and more rapidly respond to emerging diseases.”

Additionally, Althoff’s group will develop outreach and educational activities dedicated to increasing participation of historically underrepresented groups in computer science education, research and careers. The group also plans to share its results through public open-source software and workshops.

“I am honored to receive this prestigious award and am thankful to the NSF and U.S. taxpayers for supporting our research,” Althoff said. “We will use this award funding to positively impact U.S. health and well-being.”

Leilani Battle: Taming troves of data to gain actionable insights 

With troves of data at our fingertips, understanding and presenting that information requires more than just a good eye. As co-director of the UW Interactive Data Lab, Leilani Battle is helping to visualize the future of big data. She earned a CAREER Award for “Behavior-driven testing of big data exploration tools,” a proposal rich with insight into the future of data visualization. 

“Throughout my research I interweave core principles from HCI and visualization with optimization and benchmarking techniques from data management,” she said. “My CAREER award is the culmination of those ideas.”

Currently, evaluating the tools used to glean these analyses remains challenging, in no small part due to sheer volume. Used for everything from determining how climate change is addressed to which investments to prioritize, these computational tools boast an imposing number of use cases, making standardization and benchmarking a difficult task.

Battle’s project seeks to streamline the process. A key aim targets the development of an automated testing software that can determine the capability of a data exploration tool. Another focuses on discovering the types of problems these tools and evaluation methods may introduce along the way. 

Her team will also work with leaders in the visualization community, both in industry and academia, to optimize the software’s performance, and will develop programs that help students learn fundamental visualization and research skills.

Battle was also named a 2023 Sloan Fellow earlier this year. “It’s an honor to be recognized through these awards for the kind of work that I do,” she said. 

Simon Shaolei Du: Helping over-parameterized models go mainstream

Though Simon Shaolei Du focuses on the theoretical foundations of artificial intelligence, his findings have several applications in the real world. His proposal “Toward a foundation of over-parameterization” underlines the power of over-parameterized models and enumerates methods to help them go mainstream. These models have the ability to revolutionize several domains, including computer vision, natural language processing (NLP) and robotics. But they remain resource-intensive, costing millions to train and implement effectively. 

Support from the CAREER Award will help Du and his team design a resource-efficient framework to make modern machine learning technologies more accessible, better optimized and easier to evaluate. From both a theory and a practice standpoint, the project also aims to make measurable gains — characterizing the properties of over-parameterization while also deploying algorithms on real-world applications. 

By gaining a deeper understanding of over-parameterization, its benefits and drawbacks, Du hopes to form a mathematical theory that rigorously characterizes the optimization and generalization properties of neural networks. 

“Deep learning technology has been very successful in practice but its theoretical understanding is still limited,” Du said. “One prominent feature that distinguishes neural nets from previous machine learning models is over-parameterization — neural nets use many more parameters than what is needed. I found this phenomenon of neural nets very interesting as it requires a radically different theory to understand.”

Kevin Jamieson: Closing the loop on a new paradigm in machine learning

Are you aware that every time you scroll through social media or binge-watch on a streaming platform, you’re participating in an adaptive experimental design? The algorithms behind these platforms are constantly testing hypotheses and using your engagement as evidence. With machine learning, these systems curate content based on your preferences and behavior, and the success of their design is directly linked to the accuracy of their recommendations.

It’s one example of closed-loop learning, a learning paradigm that Kevin Jamieson, who earned his bachelor’s from the UW Department of Electrical & Computer Engineering and is now the Guestrin Endowed Professor in Artificial Intelligence and Machine Learning at the Allen School, explores in his NSF CAREER proposal. In “Non-asymptotic, instance-optimal closed-loop learning,” he illustrates the practical impact of harnessing closed-loop data collection strategies. While existing efforts in data collection can help provide recommendations for one’s entertainment feed, for instance, other areas — especially those in science and healthcare — have proven more imposing, both in terms of time and cost. As a result, closed-loop strategies have not yet been widely adopted in places such as medical labs, due to lower predictability and higher stakes. 

With his CAREER Award project, Jamieson seeks to crystallize closed-loop constructs into effective and reliable data-collection strategies. One instance of this playing out, he suggests, deals with clinical drug trials. As the algorithm adapts, the time needed to identify a cure could decrease significantly. Ambitious yet applicable, it’s an insight that could save lives.

“When a patient walks into the clinic with a particular health state, taking actions like running tests can help determine what the true state of health is, and other actions like prescribing medication can move the state towards more favorable states and outcomes,” Jamieson said. “My work aims to minimize the number of patients and actions required to learn the optimal way to act in this environment.”

Jeff Nivala: Pursuing new information architectures based on synthetic polymers   

For Jeff Nivala, noticing the little things is something of second nature. 

When he was a Bioengineering undergraduate at UW, Nivala took part in the Genetically Engineered Machines Competition (iGEM), an international contest that tasked teams with building “the coolest, most world-changing technologies” through synthetic biology. But beyond the sights of the competition, something else caught his eye: iGEM’s logo. 

With a green biological cell and a mechanical gear intertwining among its letters, the emblem exemplified the field’s potential to the budding scientist. 

“This visual struck a chord with me that we can engineer and program things at the molecular level using biology,” he said. “This concept is really cool and we are only at the beginning stages of its development.”

Now a professor in the Allen School, Nivala is also co-director of the Molecular Information Systems Lab. He earned an NSF CAREER Award for “Machine-guided design of enzymatically-synthesized polymers optimized for digital information storage,” which explores the growing field of synthetic biology and how scientists can leverage the advantages of molecular information storage. Some of these pros include high densities, long shelf lives and low energy costs. 

But challenges remain. Reading and writing the data, for example, is costly when done at the molecular level. Nivala’s research seeks to address this concern by developing a new information storage medium based on synthetic polymers — a scalable solution at a lower cost.

Moreover, inexpensive, laptop-powered nanopore readers can quickly decode the polymers. Featuring lower latency and higher throughput, the system stands as an innovative alternative to traditional mass spectrometry. 

Noticing the little things, it turns out, can have big consequences. The result, Nivala said, may be a brighter future. 

“Biology and technology are both incredible in their own ways,” Nivala said. “Biology has created amazing things — just look at the living things outside your window or your own body. And then there are the incredible advancements humanity has made, especially in the age of silicon and modern electronic computing. But what if we could combine these two forces? This is the future that I hope my research can contribute to, even if it’s just a small nudge in this direction.”

With support from the CAREER Award, Nivala will also create a special topics course focused on molecular computing and information storage. Looking back — and beyond — Nivala said, provides perspective and illustrates the promise realized in the present, moment by moment, molecule by molecule. 

“I am truly humbled and honored to have been awarded the NSF CAREER Award,” Nivala said. “Even now, as I devote much of my time to mentoring and guiding the next generation of scientists, I am constantly learning from more experienced colleagues that I work with here at UW. So receiving this award is a time of reflection for me, as I am reminded of the many individuals who have contributed to my success. At the same time, it is an exciting opportunity to look forward to the future and the exciting work that lies ahead.”

Chris Thachuk: Programming molecules and advancing synthetic biology

Chris Thachuk bridges the gap between computer science and synthetic biology, focusing on creating programmable matter at the molecular level. He earned a CAREER Award for “Facile molecular computation and diagnostics via fast, robust, and reconfigurable DNA circuits,” an exercise in turning the stuff of science fiction into a reimagination of what can be made real. 

In his proposal, Thachuk asks if we can envision a future where “smart,” programmable molecules process information and manipulate matter in the unseen world. From the 30-ton ENIAC — about the weight of three adult elephants — to the wearable devices of today, computers have gotten progressively smaller as technology has advanced. 

Thachuk seeks to create new design principles and architectures centered around robust “field-programmable” DNA circuits, which will be reconfigurable by non-experts without sophisticated equipment. Dubbed “DNA strand displacement (DSD) architectures,” these tiny testaments to the imagination carry an abundance of possibilities for a number of fields, including global health, diagnostics, environmental monitoring, molecular manufacturing, and more.

These architectures will be put to the test next year in an undergraduate molecular computation course to be offered by Thachuk and supported by this award. As part of the course, Allen School students, with no assumed wet lab experience, will design, build and experimentally validate state-of-the-art molecular circuits and other nanoscale devices. 

For Thachuk, it’s an elegant pairing between technical goals and educational outcomes. 

“Ideas from computer science will be the driving force behind future progress in programmable control at the nanoscale and in similarly complex environments,” Thachuk said. “The best way to accelerate that path is by exposing CSE students to these broader perspectives on computing early in their training through undergraduate classes that explore the fuzzy boundaries of disciplines.”

Yulia Tsvetkov: Harnessing the power of language against online bias

When Yulia Tsvetkov started reading social science literature about gender bias, she realized some of her personal experiences were not unique to her. 

“I also realized that as a natural language processing and machine learning researcher, I have a constructive and powerful way to combat such discrimination,” she said, “while solving interesting scientific and technical problems that I’m passionate about.”

Now an NSF CAREER Award and Sloan Research Fellowship recipient, Tsvetkov is using her talents to help others fight bias online. With “Language technologies against the language of social discrimination,” she targets a growing phenomenon in how we interact on the internet. While the growth of social networking services has produced several benefits, it’s also provided a medium for implicit bias to enter, often unimpeded.

Moreover, the effects of bias, Tsvetkov outlines, are manifold — each being harmful in nature. 

At UW, her research group focuses on developing practical solutions to NLP problems and understanding how language online shapes or is shaped by societal conditions. Tsvetkov’s CAREER Award project aims to create NLP technologies to combat societal biases that find their way into the collective discourse.

The developed models will be able to detect and counteract implicit bias and harmful speech. Her work also involves building new methods to interpret these deep-learning models, with the end goal of creating a safer and more civil cyberspace. 

“If we are able to detect such language at scale, there are opportunities to empower people and prevent them from being the targets of discrimination,” she said. “I’m excited about this topic because in addition to its potential social impact, there are exciting and timely technical challenges we’ll need to solve.”

Amy Zhang: Empowering vulnerable users online, with a little help from their friends

Sometimes we need a little help from our friends. At least, that’s the idea behind Squadbox, a crowdsourcing tool that Amy Zhang developed while a doctoral student at MIT. Squadbox assembled a “squad” of capeless crusaders (friends) who would use their powers of familiarity and solidarity to filter messages and safeguard users from attacks. Who needs a shield made of vibranium when you have a built-in support system? 

As director of the Social Futures Lab at the Allen School, Zhang is continuing to unite in order to fight online harassment. In “Tools for user and community-led social media curation,” she outlines methods for creating a safer and healthier online environment. Users today, especially those from marginalized communities, are at the mercy of a platform’s content filters, which often fail at distinguishing the harmful from the benign. This phenomenon can create a bottleneck of homogenized content, wherein the most vulnerable are left without recourse. 

Zhang’s research seeks to solve this problem. Using a mixed-methods approach, her work will gather information from a range of users facing the greatest challenges under current designs, including marginalized individuals who are often targets of harassment, neurodiverse users, journalists and content creators, among others. 

Another goal involves lowering the barrier to entry. By building collaborative systems and consulting user feedback, Zhang’s research will allow users to curate their platforms without the need for technical expertise or a large time commitment. One project in the works, for example, has users designing their own filters to keep platforms safe and accessible. 

“I am honored to receive early career funding for this proposal and have the opportunity to impact the future of social technology, particularly given the quickly changing landscape of social media today,” Zhang said. “Through my research, I look forward to introducing new tools into that landscape that empower people and communities to take control of their online social environments and that better serve the needs of marginalized users. I am excited that this funding will additionally support education and outreach to engage the next generation of computer scientists and the public on the social ramifications of technology design.” Read more →