NLP Exercises
A collection of coding practice using common NLP toolkits and libraries.
As part my masters in computer science at Portland State University, I took all the Natural Language Processing courses offered. After taking various other machine learning courses in the hopes of learning more, I kept finding that NLP was typically only briefly mentioned. So I decided to design my own NLP course. For Winter and Spring terms 2023, I performed independent research under Dr. Ameeta Agrawal’s supervision.
A major motivation of this work is to get hands-on coding experience implementing the tools and libraries I have learned about. This practice work is collected in the /implementations dir. As a kinetic learner, this really solidifies the topics in my mind.
Additionally my primary topic of research is multilingual models, and specifically how they can increase the performance of low resource languages. My motivation for this topic comes from my bachelor’s, where I primarily studied foreign languages, with a minor in computer science. There were plenty of online resources to assist my studies in French, Spanish, and German. However, I could find hardly any content for Wolof which I learned while studying abroad in Senegal. Similarly in studying NLP there is a dearth of NLP tools for African languages. It’s important to me that everyone has access to similar ML and NLP capabilities. Therefore I want to use my skills with language, NLP, and programming, to expand the resources available to low resource languages like Wolof.
Toward that end I am reading current research papers on the topics of large language models, multilingual models, expansion of resources for low resource languages, transfer learning, and few-shot learning. Of course I have also added some exercises and research on ChatGPT and Prompt Engineering, as it was released during this time.
I will finish off these studies with a final project and research paper in June of 2023. See the /project dir for more information.