Angeline Aguinaldo

My work is about finding the right unifying language: representations that make structure visible and transferable, from robot task plans to LLM agent systems.

Angeline Aguinaldo

About

I am a senior research scientist at the Johns Hopkins University Applied Physics Laboratory, where I work on agentic reasoning, planning, and safe autonomous behavior for AI-controlled systems. I recently completed my Ph.D. in Computer Science at the University of Maryland, College Park, advised by William Regli. My dissertation, Sequential, Hierarchical, and Analogical Plan Transfer in Robotics, received the Larry S. Davis Doctoral Dissertation Award.

My research sits at the intersection of AI planning, knowledge representation, and category theory. I build formal frameworks for how agents can reuse and adapt task strategies across domains without retraining, grounded in category theory and validated on both robotic planning benchmarks and LLM agent systems. At JHU APL, I lead research on LLM-based agent architectures for long-horizon planning, structured world modeling and policy evaluation under uncertainty, and safety-critical autonomous systems.

News

Research

Compositional plan transfer and cross-domain generalization.

How can an agent (robotic or LLM-based) reuse learned task plans across different domains without retraining? My dissertation developed three categories of plan transfer (sequential, hierarchical, analogical) with formal validation methods. More recently, I have extended this to LLM agents operating over structured APIs: a translation functor between OpenAPI schema categories satisfies the functor law, making cross-API plan transfer an algebraic guarantee rather than an empirical regularity.

Structured world modeling and planning under uncertainty.

How should a planning agent represent the world it operates in, and how should that representation behave under actions and partial observability? I have built simulation environments with probabilistic state dynamics and batch policy evaluation for sequential decision-making under uncertainty. My broader interest is in world representations that are compositional by construction: categorical state spaces where valid transitions are typed, and where the structure of the world is preserved (or deliberately transformed) across abstraction levels. This connects formal planning semantics to the architecture of learned world models.

Safe and interpretable autonomous behavior.

How do we specify what autonomous systems should and should not do, in ways that are interpretable, formally analyzable, and robust to changes in the system or environment? At JHU APL I have built formal representations of behavior for safety-critical autonomous systems, most recently in a NASA Technical Memorandum (NASA/TM-20250010771, 2025) with NASA Langley applying category-theoretic systems ontology to aviation autonomy integration.

Experience

  • 2017 – now Johns Hopkins University Applied Physics Laboratory, Senior Research Scientist
  • 2018 – 2025 University of Maryland, College Park, Ph.D., Computer Science
  • 2022 – 2023 Topos Institute, Research Associate
  • 2012 – 2017 Drexel University, B.S. & M.S., Biomedical and Electrical Engineering

Awards & Honors

  • 2026 Special Achievement Award · Johns Hopkins University Applied Physics Laboratory
  • 2025 Larry S. Davis Doctoral Dissertation Award · UMD Department of Computer Science
  • 2023 Best Paper Award · AAAI Fall Symposium on Unifying Representations for Robot Application Development
  • 2018 – 2021 Special Achievement Award · Johns Hopkins University Applied Physics Laboratory

Dissertation

Aguinaldo, A. Sequential, Hierarchical, and Analogical Plan Transfer in Robotics. Ph.D. Dissertation, University of Maryland, College Park, 2025. Advisor: William Regli.

Full text (DRUM) Defense slides

Publications

Peer-Reviewed Articles

  1. Aguinaldo, A., Patterson, E., Regli, W. Automating Transfer of Robot Task Plans Using Functorial Data Migrations. IEEE Transactions on Automation Science and Engineering, 2025. paper
  2. Aguinaldo, A., Patterson, E., Fairbanks, J., Regli, W., Ruiz, J. A Categorical Representation Language and Computational System for Knowledge-Based Planning. AAAI Fall Symposium on Unifying Representations for Robot Application Development (URRAD), 2023.Best Paper paper slides
  3. Player, R. A., Aguinaldo, A., Merritt, B. B., et al. The META Tool Optimizes Metagenomic Analyses Across Sequencing Platforms and Classifiers. Frontiers in Bioinformatics, 2:969247, 2023. paper
  4. Aguinaldo, A., Bunker, J., Pollard, B., Shukla, A., Canedo, A., Quiros, G., Regli, W. RoboCat: A Category Theoretic Framework for Robotic Interoperability Using Goal-Oriented Programming. IEEE Transactions on Automation Science and Engineering, 2021. paper
  5. Winter, M., Mankowski, W., Wait, E., De La Hoz, E. C., Aguinaldo, A., Cohen, A. R. Separating Touching Cells Using Pixel Replicated Elliptical Shape Models. IEEE Transactions on Medical Imaging, 2019. paper
  6. Caino, M., Seo, J., Aguinaldo, A., et al. A neuronal network of mitochondrial dynamics regulates metastasis. Nature Communications, 2016. paper

Technical Reports

  1. Mahlum, M., Jarvis, S., Niu, N., Aguinaldo, A., Hicks, A., Levitt, I. Formal Structures in Systems Ontology towards Air Traffic Management Architectures. NASA Technical Memorandum NASA/TM-20250010771, November 2025. paper

Workshop Papers & Preprints

  1. Aguinaldo, A. Functorial Plan Transfer: Compositional Generalization for LLM Agents via OpenAPI Schema Categories. In preparation, 2026.
  2. Aguinaldo, A. Reinforcement Learning with Verified Functorial Rewards. In preparation, 2026.
  3. Aguinaldo, A., Regli, W. Modeling traceability, change information, and synthesis in autonomous system design using symmetric delta lenses. ICRA Compositional Robotics Workshop, 2022. paper
  4. Aguinaldo, A., Regli, W. Encoding Compositionality in Classical Planning Solutions. IJCAI Workshop on Generalization in Planning, 2021. paper
  5. Aguinaldo, A., Regli, W. A Graphical Model-Based Representation for Classical AI Plans using Category Theory. ICAPS Workshop on Explainable AI Planning, 2021. paper
  6. Aguinaldo, A., Chiang, P., Gain, A., Patil, A., Pearson, K., Feizi, S. Compressing GANs using Knowledge Distillation. arXiv preprint, 2019. paper

Books

  1. Co-author. Relational Thinking: From Abstractions to Applications. Digital textbook, Topos Institute, 2024. read online

Selected Talks

  • 2025 Analogical Plan Transfer in Robotics using Functorial Data Migrations. AMS Joint Mathematics Meetings, Applied Category Theory Special Session. slides
  • 2024 "What's the plan?", asked my robot. JuliaCon. video slides
  • 2023 A Category Theoretic Approach to Planning in a Complex World. Microsoft Future Leaders in Robotics and AI Seminar Series. video
  • 2022 Category theory for automated planning and program compilation in robotics. Topos Institute Berkeley Seminar. video
  • 2022 Contextual affordances in context-aware autonomous systems. NIST Compositional Structures for Systems Engineering and Design Workshop.
  • 2022 Modeling traceability, change information, and synthesis in autonomous system design using symmetric delta lenses. ICRA Compositional Robotics Workshop.
  • 2022 Applications of category theory to automated planning and program compilation in robotics. Xerox PARC Design Seminar.

Full list of talks on the CV.

Writing

Software

AlgebraicJulia: open-source contributor to the ecosystem of Julia packages for applied category theory, including Catlab.jl and AlgebraicRewriting.jl, used for categorical databases, scientific computing, and structured knowledge representation.

string-diagram-viewer: interactive tool for visualizing string diagrams from symmetric monoidal categories.

Additional code on GitHub.

Service & Community

  • 2026Co-organizer, Applied Category Theory Special Session, AMS Joint Mathematics Meetings.
  • 2024 – 26Seminar Series Chair, Agentic AI and Human-AI Collaboration, Johns Hopkins University Applied Physics Laboratory.
  • 2023Local Organizer, Applied Category Theory (ACT) Conference.
  • 2023Co-organizer, ICRA Compositional Robotics: Mathematics and Tools Workshop.
  • 2022 – 23Co-organizer, The Adjoint School.
  • 2021Student, The Adjoint School.
  • ongoingReviewer, International Conference on Automated Planning and Scheduling (ICAPS).
  • ongoingReviewer, Mathematical Structures in Computer Science (MSCS).

In the Media

  • 2025 UMD Computer Science · Larry S. Davis Doctoral Dissertation Award recipient.
  • 2023 UMD Computer Science · Best Paper Award at AAAI Fall Symposium.
  • 2020 JHU APL · HADR program named a 2020 Fast Company World Changing Ideas Finalist in the AI & Data category.
  • 2019 JHU APL · APL Shaping an Intelligent Approach to Disaster Response and Relief.