Five young researchers receive NSF CAREER Awards

• Kaize Ding, an assistant professor of statistics and data science at the Weinberg College of Arts and Sciences, will receive $599,999 over five years from the NSF’s Division of Information and Intelligent Systems.  

• Duri Long, an assistant professor of communications at the School of Communications, will receive $545,796 over five years from the NSF’s Division of Information and Intelligent Systems.  

• Feng Ruan, an assistant professor of statistics and data science at Weinberg, will receive $416,950 over five years from the NSF’s Division of Mathematical Sciences.  

• Lisa Volpatti, an assistant professor of biomedical engineering and chemical and biological engineering at the McCormick School of Engineering, will receive $750,000 over five years from the NSF’s Division of Materials Research.
• Jakub Witaszek, an assistant professor of mathematics at Weinberg, will receive $400,000 over five years from the NSF’s Division of Mathematical Sciences.  

Ding studies data mining, machine learning and large foundation models. His recent research focus is to develop reliable and efficient AI systems for autonomous decision-making, with applications in areas including healthcare, environment computing and scientific discovery.

With his CAREER award, “OPENALIGN: Towards Open-World Preference Alignment for Large Language Models,” Ding will develop new methods to help large language models (LLMs) remain aligned with human preferences and values in “open-world” environments, or situations that differ from the controlled conditions and abundant, high-quality data typically assumed during AI development.

Today’s AI alignment methods often depend on large amounts of carefully labeled human-feedback data. But when AI systems encounter a new task, a specialized field or an unfamiliar situation, such data may be scarce, unreliable or constantly changing. Ding’s project will develop methods that allow LLMs to learn human preferences using substantially less supervision, while making them more robust to challenges such as noisy feedback, previously unseen data and shifts in human preferences over time.

“AI systems increasingly operate in environments that are much more complex and dynamic than the settings in which they were originally trained,” Ding said. “Our goal is to develop AI alignment methods that remain data-efficient, reliable and adaptable when models encounter new tasks, new domains and evolving human needs.”

Beyond its research contributions, the project will integrate AI alignment research into education and workforce development. Ding will develop a new graduate course on data-centric AI, mentor undergraduate and graduate researchers and create outreach activities that introduce K–12 students and broader audiences to AI alignment and responsible AI development.

Long investigates how to support human agency and critical decision-making surrounding AI by both fostering public AI literacy and designing AI-powered creativity support tools. With her CAREER award, “Generative Friction: A New Paradigm for the Design of AI Assistants to Foster Creative Reflective Practice,” Long will explore how to design generative AI tools to support artists’ creative processes. The tools will provide opportunities for creative reflection by critiquing or reframing artists’ work, making the process of creating more challenging, or offering surprising ideas. The resulting software will be open source.

“The creative process is rarely linear,”  Long said. “Oftentimes creative inspiration comes from working through unexpected problems, exploring unfamiliar materials, or learning from feedback and critique. I’m interested in building AI tools that leverage these kinds of friction to keep artists at the forefront of the creative process, rather than offloading it to generative models.”

Long also will develop new curricula to build AI literacy among students pursuing creative careers. Artists will use the new generative AI tools during a Northwestern-hosted residency, resulting in a public gallery show of artworks at the end of the project. This show will expand public discourse around AI use in creative fields. 

Ruan studies the statistical and algorithmic foundations of feature learning, using tools from high-dimensional statistics and optimization to understand how models learn informative representations from data.

With his CAREER award, “Beyond Multi-Index Models: Statistical and Algorithmic Foundations for Feature Learning in Compositional Architectures,” Ruan will study how machine learning and AI systems learn useful features from high-dimensional data. Building on multi-index models, he will extend these ideas to richer compositional architectures resembling modern neural networks, with the goal of developing more robust, data-efficient methods that require less problem-specific tuning.

Ruan’s team will disseminate new methods through open-source software and interdisciplinary collaborations. 

Ruan’s project will include mentoring undergraduate and graduate students and helping develop a new course in statistics and machine learning at Northwestern University. 

Volpatti explores ways to improve the safety and efficacy of immunotherapies by reshaping how cells deliver and experience immune signals. Along with her team, Volpatti engineers proteins and nanomaterials to limit side effects and expand the treatment of immunotherapy drugs. 

With her CAREER award, “Structure-Function Analysis of PEG-Lipid Shedding for Rational Design of Targeted Nanoparticles,” Volpatti will study surface components of nanoparticles to understand how a material’s structure determines its stability. The project will attach nanoparticles with precisely defined surface components to track how the components move and detach from the nanoparticle overtime. This research will help scientists create nanoparticles with surfaces that remain stable and functional in different scenarios. 

“Our lab is interested in understanding what happens to nanomedicines after we put them into complex biological environments, where their surfaces are dynamic rather than static,” Volpatti said. “Inspired by findings from our lab’s first paper, this project will examine how PEG-lipids behave across both polymeric and lipid-based nanoparticles and uncover how molecular structure and nanoparticle architecture together control their retention or shedding. We are incredibly grateful for the opportunity to build on this work and establish design principles that can ultimately help us engineer nanomedicines with more predictable behavior in the body.”

Volpatti also will create an education program with local community colleges, in collaboration with the University of Chicago Pritzker School of Molecular Engineering, to train materials science and engineering students. The partnerships will include hands-on research, mentoring and scientific programming. 

Witaszek studies the intersection of algebraic geometry, commutative algebra and arithmetic geometry. 

With his CAREER award, “Higher Singularities and Birational Geometry in Positive and Mixed Characteristics,” Witaszek will develop new mathematical methods and theories to further understanding of the shapes described by natural equations called polynomial equations. These shapes have applications in various fields, such as computer science, biology and robotics. 

Witaszek will organize a series of workshops, conferences and other events, including some specifically for students and early career mathematicians. In these events, he will include education on advancements in AI to familiarize students with AI tools.