- 2019-- : Applying Science at the Microsoft Translator team.
Disclaimer: **Nothing** in my profile, repositories, commits, posts, or responses is related to or represents the interests or opinions of any of my employers unless otherwise **explicitly** noted.
| Year | Title | Tags | Authors | Links |
|---|---|---|---|---|
| 2021 | Adapting MARBERT for Improved Arabic Dialect Identification: Submission to the NADI 2021 Shared Task | EACL, WANLP (2021); Dialect Classification; SOTA |
Badr AlKhamissi, Mohamed Gabr, Muhammad ElNokrashy, Khaled Essam | conf., RG |
| 2020 | Deep Diacritization: Efficient Hierarchical Recurrence for Improved Arabic Diacritization | COLING, WANLP (2020); Diacritization; SOTA |
Badr AlKhamissi, Muhammad N. ElNokrashy, Mohamed Gabr | conf., arXiv, code, PWC |
| 2020 | Score Combination for Improved Parallel Corpus Filtering for Low Resource Conditions | WMT (2020); Low Resource |
Muhammad N. ElNokrashy, Amr Hendy, Mohamed Abdelghaffar, Mohamed Afify, Ahmed Tawfik, Hany Hassan Awadalla | conf. |
- Meta learning
- Data filtering
- Curriculum learning
- Classical (nimble) expert systems
- Modern (slow, lumbering) machine learning systems
- Combining both
- Sequence/behavior modeling
- Adding refinement typing and logical typing to hobby language
- Writing compiler of hobby language