LAB 237 - Open AI Research Lab
Welcome to LAB 237, an independent, research-driven initiative advancing Artificial Intelligence (AI), Natural Language Processing (NLP), speech technology, data science, and open research resources. We build datasets, benchmarks, and practical AI tools addressing research challenges in Cameroon, across Africa, and beyond.
🎯 Our Mission
- Language Documentation: Support the documentation, digitization, and computational study of Cameroonian languages and language varieties.
- Open Dataset Development: Build carefully documented datasets and benchmarks for low-resource NLP, speech technology, and other AI applications.
- Research and Evaluation: Develop reproducible methods for evaluating language models and speech systems in underrepresented linguistic settings.
- Accessible AI: Explore efficient models, including small language models (SLMs), and practical AI tools suited to local needs and resource constraints.
- Community Collaboration: Create opportunities for researchers, students, language speakers, and other collaborators to participate in responsible AI research.
📦 Research Areas and Projects
Language and Speech Technology
- CamSpokenNLI: A research resource for spoken Natural Language Inference (NLI) in Cameroon Pidgin English and Camfranglais, investigating semantic reasoning, cross-variety transfer, and speech-recognition challenges.
- Cameroonian Speech Resources: Research into speech data, transcription, and evaluation methods for underrepresented Cameroonian languages and varieties.
Data-Centric AI and Benchmarks
- Localized Text-to-SQL: Developing datasets and evaluation methods for translating locally relevant natural-language queries into executable SQL.
- Cameroonian Document AI: Exploring datasets and benchmarks for document understanding, optical character recognition (OCR), and related visual-language tasks.
These projects represent our research directions. Individual project repositories provide their current scope, release status, documentation, and licensing terms.
🔬 Research Principles
Our work prioritizes:
- Reproducibility: Documenting data collection, processing, experimental settings, and evaluation procedures.
- Data Quality: Using transparent annotation guidelines, validation procedures, and documented dataset limitations.
- Responsible Research: Respecting participant consent, privacy, cultural context, intellectual property, and applicable ethical requirements.
- Transparent Attribution: Documenting the roles of dataset creators, annotators, speakers, reviewers, and other contributors where appropriate.
- Honest Reporting: Distinguishing established findings from exploratory results and clearly stating the limitations of our resources.
⚖️ Licensing and Data Use
LAB 237 supports open and reproducible research. However, licensing is determined at the level of each repository and resource.
- Software and code may be released under the Apache License 2.0 where explicitly specified.
- Datasets, audio recordings, annotations, models, and documentation may have different licenses or access conditions depending on their content, permissions, consent requirements, and intended use.
- The license and terms stated in an individual repository take precedence over this general overview.
- Public availability does not automatically mean that every resource may be redistributed or used commercially.
Please consult each project's documentation before using, modifying, or redistributing its materials.
🤝 Collaborate with LAB 237
We welcome collaboration from researchers, lecturers, students, engineers, computational linguists, data scientists, and members of language communities interested in building responsible AI resources for Cameroon and other low-resource settings.
Contributions may include:
- Research design and literature review
- Data collection and language documentation
- Speech recording and transcription
- Data annotation and quality validation
- Software and model development
- Evaluation, documentation, and reproducibility checks
Participation arrangements, consent requirements, compensation, and expected forms of credit should be communicated clearly for each project.
👤 Leadership and Contact
Founder & Lead Architect: Tiani Pekins Ebika
Affiliation: Department of Computer Engineering, University of Buea, Cameroon.
Email: tiani@lab237.0rg
GitHub: https://github.com/lab-237
Hugging Face: https://huggingface.co/lab-237
Website: https://lab237.org
Explore our repositories for project documentation, datasets, code, training scripts, and evaluation pipelines.
LAB 237 - Building open, reproducible, and locally grounded AI research resources for Cameroon.