Organisation Name
Addis Ababa Science and Technology University
Industry
Education
Organisation Website
Country
Ethiopia
Sustainable Development Goals (SDGs)
SDG 2: Zero Hunger
SDG 3: Good Health and Well-being
SDG 11: Sustainable Cities and Communities
SDG 13: Climate Action
General Description of the AI tool
AksumAI is an inclusive AI platform that delivers real-time disaster alerts and health advisories to vulnerable communities in Ethiopia and East Africa via low-bandwidth voice interfaces. Using fine-tuned NLP models for Amharic, Oromo, and Tigrinya, it translates satellite flood/drought predictions and epidemic warnings into actionable voice/SMS messages—accessible even to illiterate users. By fusing local language processing with climate/health data, we reduce disaster response time by 70%, empower smallholder farmers with crop-saving insights, and enable NGOs to coordinate aid efficiently.
Github, open data repository
https://zenodo.org/records/5504175
Relevant Research and Publications
0. Girma Neshir Alemneh, Andreas Rauber, Solomon Atnafu (2020). BERT Fine-Tuning for Amharic Sentiment Classification. Workshop co-located with the Eighth Swedish Language Technology Conference (SLTC).
1. Girma Neshir Alemneh, Andreas Rauber, Solomon Atnafu (2021). Meta-Learner for Amharic Sentiment Classification. Applied Sciences, 11(18), 8489. doi: 10.3390/app11188489.
2. Girma Neshir Alemneh, Andreas Rauber, Solomon Atnafu (2021). Topic Modeling for Amharic User Generated Texts. Information, 12(10), 401. doi: 10.3390/info12100401.
3. Girma Neshir Alemneh, Andreas Rauber, Solomon Atnafu (2019). Dictionary-Based Amharic Sentiment Lexicon Generation. In Information and Communication Technology for Development for Africa (pp. 311-326). Springer International Publishing.
4. Girma Neshir Alemneh, Andreas Rauber, Solomon Atnafu (2020). Negation Handling for Amharic Sentiment Classification. Workshop co-located with AAAI 2020. Available at: http://kdd.cs.ksu.edu/
5. Girma Neshir Alemneh, Andreas Rauber, Solomon Atnafu (2019). Corpus-Based Amharic Sentiment Lexicon Generation. CEUR Proceedings, Forum Artificial Intelligence Research (FAIR). Available at: http://ceurws.org/Vol-2540/, and workshop co-located with AAAI. Available at: http://kdd.cs.ksu.edu/
Needs
Personnel