Problem
Research questions and friction points this paper is trying to address.
maternal mortality
on-device AI
retrieval-augmented generation
low-resource settings
clinical decision support
MAM-AI: An On-Device Medical Retrieval-Augmented Generation System for Nurses and Midwives in Zanzibar
🤖 AI Summary
This study addresses the challenge faced by nurse-midwives in Zanzibar, Tanzania, who struggle to access authoritative maternal care guidelines due to limited internet connectivity and insufficient training. To bridge this gap, the authors propose the first fully offline, on-device medical retrieval-augmented generation (RAG) system, designed to operate on standard Android devices. The system leverages EmbeddingGemma (300M) to retrieve relevant passages from a local repository of 63,650 guideline segments and employs a quantized Gemma-4B model to generate citation-grounded responses. Through tailored prompt engineering, the approach reduces the model’s refusal rate from 33% to 3%, achieves retrieval performance comparable to cloud-based systems, and attains state-of-the-art factual consistency in generated answers. Empirical evaluation demonstrates that answer quality is highly dependent on knowledge base coverage. Both the system and its evaluation framework are publicly released.
📝 Abstract
Maternal and newborn mortality remain among the highest in sub-Saharan Africa, where midwifery care is often delivered by nurses who lack midwifery training to international standards, and consulting authoritative guidance at the point of care is hard: the guidelines are long and connectivity is intermittent. We present MAM-AI, a medical question-answering assistant for nurse-midwives in Zanzibar that runs entirely on a commodity Android device: a question is embedded (EmbeddingGemma, 300M) and matched against a curated corpus of 87 guideline documents (63,650 passages), then answered with citations by a 4B int4 generator (Gemma 4 E4B), fully offline, with no query leaving the device. We evaluate the exact deployed configuration with a layered methodology -- retriever, generator under oracle context, end-to-end, and latency -- scored by LLM judges validated against physician rubrics. The evaluation relocates the hard problem. On-device retrieval is essentially solved: the 300M embedder ranks third of seven retrievers and rivals cloud systems, so the passages the system needs are usually found. The small generator is what remains in doubt: adding retrieved context does not improve its answers, and at 4B it cannot be both helpful and safe at once -- of two same-size candidates, the more helpful one commits genuine dangerous errors, so we deploy the other, which is about twice as faithful to its sources (as faithful as a frontier model), and recover its helpfulness with a redesigned prompt that cuts deflection from 33% to 3%. Corpus quality is decisive for the same reason: where the corpus holds the right passage the answer is specific and actionable, and where it does not it goes vague. MAM-AI is a thoroughly evaluated, open-source research prototype, not a fielded product; the system, knowledge base, benchmarks, and evaluation harness are released.
Innovation
Methods, ideas, or system contributions that make the work stand out.
on-device AI
retrieval-augmented generation
offline medical QA
quantized LLM
low-resource deployment
🔎 Similar Papers
No similar papers found.
