Masakhane Community Documentation and Open NLP Guides
Community documents, meeting notes, initiatives, publications and onboarding guidance for participatory NLP research on African languages.
Search titles, authors, subjects and dates from across the DARE African Knowledge Repository.
60 resources found in the repository.
Community documents, meeting notes, initiatives, publications and onboarding guidance for participatory NLP research on African languages.
Four-unit MOOC on farm data, data sharing principles, FAIR and open data, data analysis and visualization, and open licensing for data.
Accessible Colab tutorials for teaching machine learning basics, with initial versions drawn from teaching at Deep Learning Indaba and other summer schools.
Five-course series on open, locally relevant and unbiased AI training data for African languages, earth observation, data sharing standards and bias mitigation, with SCORM and video replication materials.
Computer-vision lab notebooks from Data Science Africa 2018, including image pipelines, YOLO and a CNN notebook for cassava disease diagnosis.
A practical introduction to natural-language processing with Python and the NLTK library, covering text processing, classification, tagging, parsing and corpus access.
Graduate-level open textbook on algorithmic aspects of machine learning, covering algorithms and data structures, theory of computation and applied discrete mathematics as used in learning theory.
An introduction to computational and inferential thinking for data science, covering data structures, randomness, sampling, hypothesis testing, estimation and prediction using Python.
A comprehensive textbook on artificial intelligence foundations: intelligent agents, search, constraint satisfaction, logic, reasoning under uncertainty, decision making, learning and multiagent systems.
The standard textbook on reinforcement learning: multi-armed bandits, Markov decision processes, dynamic programming, Monte Carlo methods, temporal-difference learning, policy gradients and deep reinforcement learning.
Community Hugging Face mirror of the Swahili news topic-classification corpus with train/test split.
~11.8 hours of RFI Swahili broadcast news audio, 10,180 train and 1,991 test utterances with manual transcripts at 16 kHz for ASR.