About
You are viewing the personal website of Kevin Krahn. I am a computer science student with a range of interests and programming experience.
My primary research interests are natural language processing (NLP) and computational linguistics, especially involving machine learning methods for maximizing performance in resource-constrained environments. My research to this point has focused on NLP with application to historical texts and digital humanities. Specifically, I have worked on cross-lingual transfer learning for Ancient Greek sentence embedding models as well as hierarchical language model architectures for enhancing morphological annotation for low-resource historical languages.
I also have interests in computer graphics, data visualization, game development, linguistics, ancient languages, and computer science education. My professional experience includes embedded Linux systems, full-stack web development, and mobile app development. Check my CV for more details.
Publications
-
(2024) Frederick Riemenschneider* and Kevin Krahn*. Heidelberg-Boston @ SIGTYP 2024 Shared Task: Enhancing Low-Resource Language Analysis With Character-Aware Hierarchical Transformers. In Proceedings of the 6th Workshop on Research in Computational Linguistic Typology and Multilingual NLP, pages 131–141, St. Julian's, Malta. Association for Computational Linguistics
*Equal contribution - (2023) Kevin Krahn, Derrick Tate, and Andrew C. Lamicela. 2023. Sentence Embedding Models for Ancient Greek Using Multilingual Knowledge Distillation. In Proceedings of the Ancient Language Processing Workshop, pages 13–22, Varna, Bulgaria. INCOMA Ltd., Shoumen, Bulgaria.
Selected Projects
- (2023) Ζήτησις - Semantic search tool for querying ancient texts in the original Greek text or English translations.
- (2023) Machine Translation of Ancient Greek - A neural machine translation model for translating ancient Greek to English using a fine-tuned mt5 model.
- (2023) Rust MNIST - A neural network for recognizing handwritten digits from the MNIST dataset, implemented in Rust with zero external dependencies.
- (2022) Raytracer - A somewhat basic but functional multi-threaded raytracer implemented in Rust.
- (2022) Journey Groups - A Flutter app (with Django backend) for Android and iOS for groups and individuals to track habits and other disciplines.
- (2018 - 2021) Racegame (working title) - 3D racing game for Linux and PC, written from scratch with a custom game engine, in C++ and OpenGL.
- (2017) Website for KRAHN, and outdoor furniture company with custom CMS, order management portal, and custom 3D WebGL product viewer.
- (2015) One2Three - Arithmetic puzzle game for Android.
- (2015) KABOOM! - A 3D puzzle game for PC built in 72 hours for a game jam, where it won first place.
- (2013) Double World - A 3D puzzle game for PC built in 72 hours for a game jam.