ECE594: Mathematical models of LanguageHere is a set of project ideas. This list will be evolving over the next few weeks.Low-Resource (LR)/Multilingual (ML) NLP 1. Propose a novel method for LR/ML NLP that is better than other methods (e.g., more accurate, or efficient in terms of computation or data use) in the literature. 2. An extensive analysis and comparison of a particular ML/LR prceossing task, with data-driven support for your analysis. 3. Building a language processing model for a language that does not currently have such technology. The innovation will be how you apply methods that you learned in class to that language based on its linguistic properties, data availability or language relatedness.
If you are looking for other sources of inspiration, you may find the following useful. 1. Students’ projects for COS597E (Fall 2018) 3. Stanford’s CS224N course projects
|