Two Former Graduate Students Receive Larry S. Davis Doctoral Dissertation Awards

The University of Maryland’s Department of Computer Science has announced the recipients of the 2025–26 Larry S. Davis Doctoral Dissertation Award. Angeline Aguinaldo (Ph.D. ’25, computer science), now a senior research staff member at the Johns Hopkins University Applied Physics Laboratory, and Songwei Ge (Ph.D. ’25, computer science), a research scientist at Reve, were selected for the honor. The award recognizes dissertations for their technical depth, significance, potential impact and presentation quality.

Named after Professor Emeritus Larry S. Davis, the award honors students who demonstrate innovation and excellence in computer science research. Davis, who served as department chair from 1999 to 2012, made pioneering contributions to computer vision and high-performance computing, earning an IEEE Fellowship in 1997.

Angeline Aguinaldo:

Aguinaldo’s research, advised by Professor William Regli, focuses on helping robots safely and intelligently adapt to new environments by reusing what they’ve already learned. Her work applies category-theoretic approaches to knowledge-based task planning, plan transfer and plan composition in AI and robotics.

“My goal is to help robots reuse what they’ve already learned so they can adapt safely and intelligently to new environments,” Aguinaldo said. “I do this by exploring category-theoretic approaches to knowledge-based task planning, plan transfer and plan composition in AI and robotics.”

She explained that her research offers a principled foundation for transferring task-level plans across complex, knowledge-rich environments—an important step toward creating more adaptable and trustworthy AI systems.

“As robotics and AI evolve, the need for systems that can learn, adapt and transfer knowledge across tasks is becoming increasingly critical,” she said. “By showing how category-theoretic principles can meet these needs, this research lays the groundwork for both physical robots and AI agents that can perform interdependent tasks while safely adapting to new environments with formal guarantees.”

Aguinaldo described receiving the award as an important professional milestone. 

“I am deeply honored to receive this award,” she said. “I’m grateful for the guidance and support of my advisor, Dr. William Regli, my committee members and my collaborators across the applied category theory and AI planning communities.”

She added her appreciation for the institutions that supported her work. 

“I also want to thank the faculty and staff of the UMD Department of Computer Science and the Johns Hopkins University Applied Physics Laboratory for giving me the foundation and resources to make this work possible.”

Her research, she said, has potential applications in manufacturing and mission-planning settings where privacy, safety and interoperability are essential.

Songwei Ge:

Ge’s dissertation, advised by Professor David Jacobs and Associate Professor Jia-Bin Huang, explores how generative AI systems can synthesize and augment visual content—from sketches and images to videos and 3D scenes—to enhance human creativity. 

His research continues at Reve, where he develops technologies that expand the capabilities of large-scale visual synthesis.

“I am super honored and thrilled to receive this award,” Ge said. “The recognition from the department means a great deal to me; it’s where I’ve grown as a researcher, surrounded by inspiring faculty, mentors and peers who constantly push the boundaries of computer vision and AI.”

He credited his advisors for their mentorship and guidance. 

“I could not have done this without the help of my Ph.D. advisors, David Jacobs and Jia-Bin Huang, and the department,” he said.

Ge emphasized how his research connects technology and art by empowering people to create through intuitive AI tools.

“Human creativity has always been in a symbiotic and continually evolving relationship with technological development,” he said. “Impressionism develops with synthetic pigments, film-making starts with the zoopraxiscope and video games grow with computer-generated imagery.”

Ge added that the creativity achievable with modern computing is vastly beyond imagination. He is dedicated to building machine learning models that can synthesize high-quality visual content and apply them to support human creativity, enabling anyone to create visual works using natural language or intuitive interfaces—even without formal training.

—Story by Samuel Malede Zewdu, CS Communications