My thesis on "Open Source Code and Low Resource Languages" for an MSc in Language Science and Technology at Saarland University

📖 LanMIT: A Toolkit for Improving Language Models in Low-resourced Speech Recognition based on Kaldi.

Assignment 2 for CS 11-731 Machine Translation course.

Automating healthcare QA in a noisy multilingual low-resource setting

Confidence Estimation for Black Box Automatic Speech Recognition Systems Using Lattice Recurrent Neural Networks https://arxiv.org/abs/1910.11933 or https://ieeexplore.ieee.org/document/9053264

A Text-To-Speech android app to read out Gondi articles for the Gond tribal community

Low resource machine translation using Transformers and Iterative Back translation

Fake news detection in Filipino via Multitask Transfer Learning

LOW-RESOURCE NEURAL MACHINE TRANSLATION: A BENCHMARK FOR FIVE AFRICAN LANGUAGES

Dynamic Transfer Learning for Low-Resource Neural Machine Translation

The Kinyarwanda and Kirundi Languages Toolkit (KKLTK) is a Python package for Kinyarwanda and Kirundi languages processing. KKLTK currently provides the sets of stopwords for both languages and other preprocessing tools such as Kinyarwanda and Kirundi tokenizers will be added soon. KKLTK requires Python 3.0, 3.5, 3.6, 3.7, or 3.8.

Dataset for Paper - A Neural Approach to Multilingual Sentiment Analysis in Low Resource Languages, submitted in Elsevier Expert Systems with Applications

Open-source benchmark datasets and pretrained transformer models in the Filipino language.

a repository containing the details of natural language inference dataset in Hindi

[AAAI 2021] - Simple or Complex? Learning to Predict Readability of Bengali Texts.

[EACL 2021] - Unsupervised Abstractive Summarization of Bengali Text Documents.

Train JOSA (Jopara Sentiment Analysis) corpus with traditional machine learning algorithms.

Download guarani-dominant tweets

Code for "The Greedy and Recursive Search for Morphological Productivity." Ready for use on new data.