dakotahlambert@acm.org
Assistant Professor of Computer Science at Lake Forest College
I study the structural and computational complexity of languages. Here, “structural complexity” refers to the question of how constrained the interactions among the parts are. I use techniques from universal algebra to explore this question. The related question of “computational complexity” concerns what aspects of a string must be attended to by any computational mechanism for that mechanism to be able to determine the acceptability of the string or to determine what should come next. Learning algorithms arise from appropriately decomposing and parameterizing these systems.
Particular topics of research interest include semigroups, formal logic, finite automata, computability, and, to a lesser extent, machine learning. I am also interested in functional programming and low-resource computational systems.
As I study machine learning and properties of languages, I am frequently asked about large language models (LLMs). In brief, I have less than zero interest in them. Their training and use are legal and ethical minefields. But beyond that, it’s just painful to see effective systems be replaced by incapable pseudotools that squander time and energy.
The Language Toolkit is an interactive theorem-prover for subregular logics, implemented as a Haskell library and a collection of command-line tools. I use the system for much of my research and regularly extend it to account for new discoveries. With it, you can do the following and much more:
This system is also available on Hackage.
Find more on Github.