Arabic Diacritics Reading Tutor with Automatic Tashkeel Feedback for Learners
Learners of Arabic, from schoolchildren in Pakistan and Nigeria to expatriate workers in the Gulf, struggle to read text without short vowel marks (tashkeel). Build a tutor that removes the marks from a passage, asks the learner to add them, and checks the answer against an automatic diacritiser and a verified reference, explaining errors by grammar rule. Measure diacritiser accuracy and whether learners improve.
Intermediate
A small team, or one strong student willing to learn something new.1 to 2 semesters · 2 students
7 to ship
4 optional extrasSuggested stack
What you should ship
- A collection of graded reading passages with verified diacritics, from openly licensed sources
- An automatic diacritiser trained on the Tashkeela corpus or a cleaned version, compared with an existing tool
- An exercise mode that removes marks and lets the learner add them with an easy on-screen keyboard
- Answer checking that separates word-internal marks from case endings, which depend on grammar
- Explanations for common case-ending errors, written with an Arabic teacher
- Progress tracking by error type
- An evaluation of diacritiser accuracy (diacritic and word error rates) on held-out text
If you have time left
- Text-to-speech playback of the correctly marked passage
- Adaptive difficulty based on the learner's errors
- Support for learners who speak Urdu, Hausa or English, with explanations in their language
- Speech input: the learner reads aloud and the tutor checks vowels
The problem: reading without the vowels
Most Arabic text is written without short vowel marks. Fluent readers fill them in from grammar and context; learners cannot. Students learning Arabic in schools across Pakistan, Nigeria, Indonesia and Europe, and expatriate workers in the Gulf learning to read signs and documents, find this the hardest step. Teachers explain case endings and patterns, but practice with instant, accurate feedback is rare.
Automatic diacritisation, adding the marks with software, is a well-studied NLP problem. Turning it into a practice tool for learners is less common.
What already exists and the gap
- CAMeL Tools, from the CAMeL Lab at New York University Abu Dhabi, offers Arabic NLP including morphological analysis useful for diacritisation.
- Mishkal is an open-source Arabic diacritiser.
- Language apps such as Duolingo teach Arabic, with limited focus on reading unmarked text.
The gap is a reading tutor that uses diacritisation to give feedback, separates easy errors from grammar-dependent ones and explains them.
Inspiration
The Tashkeela corpus, published by Taha Zerrouki and Amar Balla in Data in Brief (2017), provided millions of diacritised words for training and has become a standard resource. Your project uses it to train a diacritiser, then turns the model into a teaching tool. It complements the Urdu learning tools already on Sefism with an Arabic one.
What you build
- Learners practise adding marks to graded passages and get immediate, explained feedback.
- Teachers assign passages and see class-wide error patterns.
- Researchers get an evaluation of diacritisers on learner-level text.
The technically interesting part
- Two kinds of marks. Marks inside a word often follow from its form; case endings depend on syntax. Measure accuracy for each separately and be honest about the harder one.
- Trusting the reference. The model is not always right. Use verified passages for exercises and show the model's suggestion only where it is confident.
- Input design. Typing diacritics on a phone is awkward. Design an on-screen picker that is fast.
- Text normalisation. Arabic Unicode has variants for the same letters. Normalise consistently.
How to evaluate it
- Diacritic error rate and word error rate, with and without case endings, on held-out text.
- Comparison with an existing diacritiser on the same text.
- Learner study: accuracy on a pre-test and post-test after two weeks of practice, with a small group.
- Explanation quality rated by an Arabic teacher.
Suggested plan
- Weeks 1 to 3: passages, data and baselines.
- Weeks 4 to 9: train and evaluate the diacritiser.
- Weeks 10 to 14: exercise mode, checking and explanations.
- Weeks 15 to 18: progress tracking and learner study.
- Weeks 19 to 22: report and demo.
Path to a product
Schools, Arabic language institutes, and employers in the Gulf training expatriate staff all need reading practice tools. A tutor with strong feedback is a realistic edtech product. The project shows NLP and product skills for a language with a huge learner population.
Scope warning
Use only openly licensed or public domain texts and credit them. Keep explanations reviewed by a qualified teacher. Do not overclaim on case endings, where even strong models make mistakes.
Learn what you need on Sefism
- Machine learning roadmap: sequence labelling and evaluation.
- Mobile roadmap: if you build the tutor as an app.
- Hugging Face Transformers documentation: sequence models.
- PyTorch: Learn the Basics: training the diacritiser.
- Speech and Language Processing: background on tagging problems.
Get help with this idea
A diacritiser alone is not a tutor. Tauseef can help you design the exercise flow and an evaluation that measures learning, not just model accuracy. If you are new here, create a free account first.
Ideas and guidance, not finished projects
These are project ideas and scoping guidance, published free for students to use as a starting point. I do not build, write, or sell final-year projects, and I do not complete coursework for anyone. Take an idea, make it yours, and build it.
Not sure this one fits you?
I help students pick an idea that matches their skills and their deadline, then scope it down until it is finishable. Guidance only, never done for you.
Stuck on something specific?
Writing only gets you so far. If you want an answer to your situation rather than the general case, book a session and we will work through it together. Sessions are free for approved Sefism members, and a few slots open each week.
Follow along
New writing, resources and project ideas land here first.