Projects
Below are our ongoing projects for Filipino NLP. We are constantly looking for collaborators, especially those interested in contributing annotations, compute resources (credits or machines), or expertise in specific areas such as evaluation, data collection, and model post-training.
-
Developing state-of-the-art neural machine translation models for low-resource Filipino languages, focusing on Tagalog, Cebuano, and Ilocano.
-
We plan to create a high-quality instruction-tuning dataset for the four widely spoken PH languages (Tagalog, Cebuano, Ilocano, Hiligaynon), with around 10k samples per language.
Frequently Asked Questions
How are projects decided?
It's a loose process, but in general, projects are proposed and discussed by the group based on gaps we find in Filipino NLP and current research needs.
It is important that a project has a dedicated lead to drive it forward, so it also helps if your research project aligns with your own interests or ongoing work (e.g., thesis, job).
Finally, we are still an academic research group at heart, so projects that can potentially lead to publications are prioritized.
How can I join a project?
Click on any project above to see the skills needed and timeline.
If you're interested, reach out to the project lead through their contact channels with information about your background.
We welcome contributors from all levels of experience, from students to experienced researchers.
I'm not an NLP researcher, but how can I help?
There are many ways to contribute!
The best way is by donating compute resources or funding to support our work.
Compute resources can be in the form of cloud credits or sharing access to machines with GPUs.
We'll use these resources for training models and running experiments.
Also, we'll acknowledge your organization in our project pages and publications.
Finally, spreading the word about our projects and sharing our models and datasets with your networks also helps a lot!