Nathaniel Hudson, Ph.D.

A brief introduction.

prof_pic.jpg

312 John Crerar Library

5730 S Ellis Ave

Chicago, IL 60637

I am a computer scientist, currently serving as a Postdoctoral Scholar at Globus Labs out of the University of Chicago’s Department of Computer Science.

A high-level description of my research is the design of systems for serving AI on edge computing infrastructure — i.e., Edge Intelligence (EI) — for smart city applications. More specifically, my research centers around challenges related to resource limitations available at the edge for supporting EI. Trade-offs between latency, accuracy, resource usage, etc. are common themes in my work.

Some areas of study my research touches include (but are not limited to):

  • federated learning
  • service placement and request scheduling
  • lossy compression techniques
  • social mining
  • modeling of information diffusion processes
  • interdependent and complex networks
  • cyber-physical systems

recent news

Jun 30, 2025 Research that explores how active learning methods can improve the rate of novel scientific discovery in generative AI workflows has been accepted for publication at the 2025 IEEE e-Science conference. This paper specifically studies the discovery novel metal-organic frameworks (MOFs) in the MOFA workflows presented in an earlier work.
Jun 26, 2025 Our paper for Flight, a hierarchical federated learning framework, has been accepted for publication through the Future Generation Computer Systems journal.
May 20, 2025 My former summer undergraduate student mentee, Jordan Pettyjohn, was recently awarded 1st place in the ACM Student Research Competition (SRC) Grand Finals in the graduate competition. This was for his work I worked with him on investigating toxicity ablation in large language models (source, retrieved May 24, 2025).
May 16, 2025 Assistant Professorship at the Illinois Institute of Technology.
Feb 11, 2025 Paper on mitigating memorization in language models, was selected to be presented as a Spotlight Paper at this year’s ICLR conference.
Feb 7, 2025 A paper on causal discovery over hypothesis spaces has been accepted to be published in the Transactions on Machine Learning Research (TMLR). A preprint for this work can be found here.
Jan 22, 2025 Thrilled to announce a recent paper of ours investigating memorization in large language models has been accepted for publication by this year’s ICLR conference. A preprint of this work is available on arXiv and a succinct blog post on its results is also available.
Nov 21, 2024 Research awarded 1st Place in the ACM Student Research Competition at 2024 IEEE/ACM Supercomputing conference.
Oct 7, 2024 A preprint of a recent work where we explore mitigation strategies for memorization in language models has been made publicly available on arXiv. Click here for the paper. For a more brief dive into the material, please see this blog (here) on the work.
Sep 19, 2024 Very happy to announce that TaPS, an evaluation suite for execution frameworks and data management systems, has been awarded the “Best Paper Award” at the 2024 IEEE eScience conference. Read the preprint of the paper here.

selected publications

  1. “Flight: A FaaS-Based Framework for Complex and Hierarchical Federated Learning”
    Nathaniel Hudson, Valerie Hayot-Sasson, Yadu Babuji, Matt Baughman, J Gregory Pauloski, Ryan Chard, Ian Foster,  and Kyle Chard
    Future Generation Computer Systems 2025
  2. FGCS
    “QoS-aware edge AI placement and scheduling with multiple implementations in FaaS-based edge computing”
    Nathaniel HudsonHana Khamfroush, Matt Baughman, Daniel E. LucaniKyle Chard,  and Ian Foster
    Future Generation Computer Systems 2024
  3. “Smart Edge-Enabled Traffic Light Control: Improving Reward-Communication Trade-offs with Federated Reinforcement Learning”
    Nathaniel Hudson, Pratham Oza, Hana Khamfroush,  and Chantem Thidapat
    In 2022 IEEE International Conference on Smart Computing (SMARTCOMP) Jul 2022
  4. ICCCN
    “QoS-Aware Placement of Deep Learning Services on the Edge with Multiple Service Implementations”
    Nathaniel HudsonHana Khamfroush,  and Daniel E. Lucani
    In 2021 IEEE International Conference on Computer Communications and Networks (ICCCN) Big Data and Machine Learning for Networking (BDMLN) Workshop Jul 2021
  5. TNSE
    “Behavioral Information Diffusion for Opinion Maximization in Online Social Networks”
    Nathaniel Hudson,  and Hana Khamfroush
    IEEE Transactions on Network Science and Engineering (TNSE) Oct 2020