In this project seminar we will work to identify a project in machine translation, language modeling, and/or NLP, make a novel contribution, compare experimental results of the new method with the existing baseline, and discuss the results. It'll be fun.

Outline

1. Course Organization
2. Collaboration:
   • Research Organization and Collaboration Using Git
   • Overleaf
   • Doodle Polls
3. Idea and Team Formation
• Team Roles:
• Model design and implementation: most team members should be involved with this to varying degrees
• Experiments
• Error analysis, visualization, examples, website
• Literature review
• Coordination and communication
• Etc.
4. Proposal Draft
   
   • Due May 19th
   • Include the following sections:
   • Introduction, explaining what the topic is, what problem you're trying to solve (including examples), and previous works
   • Methodology, listing the techniques/algorithms/models that you think you will use and/or develop. You can change this in the future. If known, briefly describe the implementation(s)
   • Experiments, listing datasets, evaluation metrics, and how it will (or could) compare with previous works
   • Who's doing what? The "Team Roles" section above might provide you with some ideas.
   • Rough timeline
   • You can change these things in the future. This proposal is mostly to get you and your group to think about the different parts of a successful research project.
5. Implementation
   
   • Due June 23rd
6. Experiments
   
   • Due July 14th
7. Paper
   • Due Aug. 25th
   • ACL Review Forms - Important to read before and during paper-writing
   • ACL 2017 LaTeX style files

External Links

• NLP Calendar
• EMNLP 2017 Workshops
• Free Corpora:
• WMT 2017, both monolingual and bilingual data for many languages
• ACL Wiki, "Resources by Language"
• OPUS - open parallel corpora
• Corpus Processing Tools:
• https://github.com/jonsafari/habeas-corpus
• https://github.com/kpu/preprocess
• http://corpus.tools