MAM-AI: An On-Device Medical Retrieval-Augmented Generation System for Nurses and Midwives in Zanzibar

• URL: http://arxiv.org/abs/2606.29580v3
• Date: Mon, 06 Jul 2026 10:39:25 GMT
• Title: MAM-AI: An On-Device Medical Retrieval-Augmented Generation System for Nurses and Midwives in Zanzibar
• Abstract summary: We present MAM-AI, a question-answering assistant for nurse-midwives in Zanzibar.<n>A question is embedded (EmbeddingGemma, 300M) and matched against a curated corpus of 87 guideline documents.<n>A 4B int4 generator (Gemma 4 E4B) is fully offline, with no query leaving the device.
• Score: 9.0025190634417
• License: http://creativecommons.org/licenses/by/4.0/
• 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.

Related papers

• BanglaMed-QA: A Question Answering System for Healthcare Support in Bangla [0.6299766708197883]
  We introduce BanglaMed-QA, a robust QA system specifically designed for the Bangla medical domain.<n>The process begins with building a structured medical knowledge base that includes 4,493 QA pairs in 9 categories under 506 diseases.<n>The performance of the QA system has been evaluated in two aspects, with a 95% F1 score in an automated evaluation and an average human satisfaction rating of 0.9 out of 1.0.
  arXiv  Detail & Related papers  (2026-08-28T13:38:35Z)
• mamabench and mamaretrieval: Benchmarks for Evaluating Medical Retrieval-Augmented Generation in Maternal, Neonatal, and Reproductive Health [9.0025190634417]
  mamabench is a scope-filtered QA set of 25,949 items assembled from seven existing expert-authored sources.<n> mamaretrieval pairs 3,185 clinical queries with graded (0-6) relevance labels over a 63,650-chunk maternal-health guideline corpus.
  arXiv  Detail & Related papers  (2026-06-28T15:51:53Z)
• PubMed Reasoner: Dynamic Reasoning-based Retrieval for Evidence-Grounded Biomedical Question Answering [12.717987334814692]
  We introduce PubMed Reasoner, a biomedical QA agent composed of three stages.<n>Self-critic query refinement evaluates MeSH terms for coverage, alignment, and redundancy to enhance PubMed queries.<n> reflective retrieval processes articles in batches until sufficient evidence is gathered.<n>Evidence-grounded response generation produces answers with explicit citations.
  arXiv  Detail & Related papers  (2026-03-28T16:41:57Z)
• Developing and evaluating a chatbot to support maternal health care [20.297280325309924]
  We present a bot for maternal health in India developed through a partnership between academic researchers, a health tech company, a public health nonprofit, and a hospital.<n>The system combines (1) stage-aware triage, routing high-risk queries to expert templates, (2) hybrid retrieval over curated maternal/newborn guidelines, and (3) evidence-conditioned generation from an LLM.<n>Our core contribution is an evaluation workflow for high-stakes deployment under limited expert supervision.
  arXiv  Detail & Related papers  (2026-03-13T17:02:05Z)
• From Conflict to Consensus: Boosting Medical Reasoning via Multi-Round Agentic RAG [65.0203623486525]
  Large Language Models (LLMs) exhibit high reasoning capacity in medical question-answering.<n>Their tendency to produce hallucinations and outdated knowledge poses critical risks in healthcare fields.<n>Existing methods rely on noisy token-level signals and lack the multi-round refinement required for complex reasoning.
  arXiv  Detail & Related papers  (2026-02-06T08:25:30Z)
• MedForget: Hierarchy-Aware Multimodal Unlearning Testbed for Medical AI [66.0701326117134]
  MedForget is a hierarchy-aware multimodal unlearning testbed for building compliant medical AI systems.<n>We show that existing methods struggle to achieve complete, hierarchy-aware forgetting without reducing diagnostic performance.<n>We introduce a reconstruction attack that progressively adds hierarchical level context to prompts.
  arXiv  Detail & Related papers  (2025-12-10T17:55:06Z)
• MedTrust-RAG: Evidence Verification and Trust Alignment for Biomedical Question Answering [21.855579328680246]
  We propose MedTrust-Guided Iterative RAG, a framework designed to enhance factual consistency and hallucinations in medical QA.<n>First, it enforces citation-aware reasoning by requiring all generated content to be explicitly grounded in retrieved medical documents.<n>Second, it employs an iterative retrieval-verification process, where a verification agent assesses evidence adequacy.
  arXiv  Detail & Related papers  (2025-10-16T07:59:11Z)
• MedKGent: A Large Language Model Agent Framework for Constructing Temporally Evolving Medical Knowledge Graph [57.54231831309079]
  We introduce MedKGent, a framework for constructing temporally evolving medical Knowledge Graphs.<n>We simulate the emergence of biomedical knowledge via a fine-grained daily time series.<n>The resulting KG contains 156,275 entities and 2,971,384 relational triples.
  arXiv  Detail & Related papers  (2025-08-17T15:14:03Z)
• From RAG to Agentic: Validating Islamic-Medicine Responses with LLM Agents [5.391445566482165]
  Islamic texts like Avicenna's Canon of Medicine and the Prophetic Tibb-e-Nabawi encode a wealth of preventive care, nutrition, and holistic therapies.<n> Existing language-model benchmarks focus narrowly on factual recall or user preference.<n>We propose a unified evaluation pipeline, Tibbe-AG, that aligns 30 carefully curated Prophetic-medicine questions with human-verified remedies.
  arXiv  Detail & Related papers  (2025-06-18T23:14:35Z)
• LLMEval-Med: A Real-world Clinical Benchmark for Medical LLMs with Physician Validation [58.25892575437433]
  evaluating large language models (LLMs) in medicine is crucial because medical applications require high accuracy with little room for error.<n>We present LLMEval-Med, a new benchmark covering five core medical areas, including 2,996 questions created from real-world electronic health records and expert-designed clinical scenarios.
  arXiv  Detail & Related papers  (2025-06-04T15:43:14Z)
• "Knowing When You Don't Know": A Multilingual Relevance Assessment Dataset for Robust Retrieval-Augmented Generation [90.09260023184932]
  Retrieval-Augmented Generation (RAG) grounds Large Language Model (LLM) output by leveraging external knowledge sources to reduce factual hallucinations. NoMIRACL is a human-annotated dataset for evaluating LLM robustness in RAG across 18 typologically diverse languages. We measure relevance assessment using: (i) hallucination rate, measuring model tendency to hallucinate, when the answer is not present in passages in the non-relevant subset, and (ii) error rate, measuring model inaccuracy to recognize relevant passages in the relevant subset.
  arXiv  Detail & Related papers  (2023-12-18T17:18:04Z)
• MedAlign: A Clinician-Generated Dataset for Instruction Following with Electronic Medical Records [60.35217378132709]
  Large language models (LLMs) can follow natural language instructions with human-level fluency. evaluating LLMs on realistic text generation tasks for healthcare remains challenging. We introduce MedAlign, a benchmark dataset of 983 natural language instructions for EHR data.
  arXiv  Detail & Related papers  (2023-08-27T12:24:39Z)
• Medical Question Understanding and Answering with Knowledge Grounding and Semantic Self-Supervision [53.692793122749414]
  We introduce a medical question understanding and answering system with knowledge grounding and semantic self-supervision. Our system is a pipeline that first summarizes a long, medical, user-written question, using a supervised summarization loss. The system first matches the summarized user question with an FAQ from a trusted medical knowledge base, and then retrieves a fixed number of relevant sentences from the corresponding answer document.
  arXiv  Detail & Related papers  (2022-09-30T08:20:32Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.