Samuel Barham is an AI researcher at the Johns Hopkins University Applied Physics Laboratory (JHU/APL), where he brings a diversity of experience to the lab’s ongoing work in large language models (LLMs). Samuel holds a B.M. in classical piano performance and a B.A. in Russian language and literature, the latter of which was completed partly in St. Peterburg and Moscow. Following these degrees, he pursued a B.S. in computer science with a minor in mathematics, culminating in a master’s degree in artificial intelligence from the University of Maryland, College Park, where he worked under Soheil Feizi.

At the APL, Samuel Barham’s contributions have significantly shaped the landscape of the lab’s work in LLMs. Having led the development at APL of an early precursor to the LangChain framework, he displayed foresight in grasping the potential of the technology before its time. He currently acts as principal investigator (PI) for a number of APL internal research and development (IR&D) projects related to LLMs, in which he researches topics ranging from novel sampling techniques, to the benefits and risks of composable, agentive LLM design paradigms. Samuel’s role extends to numerous domains, including intelligence analysis and battlefield medicine, reflecting his commitment to practical applications that can make a real-world impact.

In his free time, Samuel works as a luthier, designing and crafting custom guitars and banjos, showcasing his creativity and artistry outside the world of AI research and academia.

Education History

  • B.S., Computer Science, University of Maryland
  • B.A., Russian Language and Literature, University of Maryland
  • B.M., Piano Performance, University of Maryland
  • M.S., Computer Science, University of Maryland

Work Experience

Associate Professional Staff, JHU Applied Physics Laboratory

Publications

Barham, S., Weller, O., Yuan, M., Murray, K., Yarmohammadi, M., Jiang, Z., Vashishtha, S., Martin, A., Liu, A., White, A. S., Boyd-Graber, J., Van Durme, B. (2023). “MegaWika: Millions of reports and their sources across 50 diverse languages.” arXiv preprint arXiv:2307.07049.

Mayfield, J., Yang, E., Lawrie, D., Barham, S., Weller, O., Mason, M., Nair, Miller, S. (2023). “Synthetic Cross-language Information Retrieval Training Data.” arXiv preprint arXiv:2305.00331.

Barham, S., Feizi, S. (2019). “Interpretable Adversarial Training for Text.” arXiv preprint arXiv:1905.12864.

Brody, J., Barham, S., Dai, Y., Maxey, C., Perlis, D., Sekora, D., Shamwell, J. (2016). “Reasoning with Grounded Self-Symbols for Human-Robot Interaction.”
In AAAI Fall Symposia Series.

Sekora, D., Barham, S., Brody, J., Perlis, D. (2017). ”Anatomy of a Task: Toward a Tentative Taxonomy of the Mind.” In AAAI Fall Symposia Series.