-
Notifications
You must be signed in to change notification settings - Fork 24
Paper Implementations
In this program, we aim to encourage the implementation of research papers to gain more experience and know-how about the different aspects of ML/NLP research. From understanding how to preprocess the data efficiently to building the model to how to set up code for proper experimentation. There are a lot of benefits in implementing papers, including the ability to digest research material faster and understanding intuitions of different components of a model.
In order to participate, make sure to add your name to this GitHub issue. Then we will organize different groups that will be assigned a particular paper. You are free to identify and select your own paper but make sure to propose it to me before moving forward.
All paper implementations must be accessible and easy to follow. We will provide reviews, feedback, and a checklist to ensure that the paper implementations are of high quality.
After identifying groups and having assigned the papers, all groups will gather biweekly and report their progress. This will allow the sharing of ideas and insights. So while you implement your own papers, you will also learn how others implement theirs and learn from that experience as well. We will provide guidance and feedback to make sure that projects are progressing. There is no deadline, we only ask that groups be responsible and report their progress even if it is minor. If additional help is needed for groups, we will work on making sure to offer that help.
When a group is finished with their implementation, they will be invited to present it to the bigger group. The completed projects will be highlighted in an upcoming issue of the NLP Newsletter, our social media websites, the main website (dair.ai), and the GitHub organization page.