Empowering Remote Healthcare with On-Premises Solar-Powered AI Units: Design and Implementation
Bibliographic record
Abstract
Rural healthcare systems face considerable obstacles such as unreliable electricity, limited internet access, and shortages of healthcare professionals, all of which impede timely medical documentation and diagnostics. This study aims to design and evaluate a solar-powered AI unit equipped with fine-tuned Large Language Models for remote clinics, enabling offline medical transcription, clinical note generation, and diagnostic support in regions with limited infrastructure. Employing a mixed-methods approach, the research combines qualitative user experience assessments with quantitative performance metrics. Four TinyLLaMA models with 1.1 billion parameters were fine-tuned to generate Subjective, Objective, Assessment, and Plan (SOAP) notes using a synthetic dataset comprising thousands of patient records and transcriptions. These models were deployed on a Raspberry Pi 5, powered by solar panels, batteries, and a Wi-Fi antenna. System performance was simulated using mockup data, with plans for validation through real-world deployment. The fine-tuned models achieved high transcription accuracy, rapid note generation, and substantial diagnostic precision on mockup data, with a balanced demographic distribution. Qualitative feedback emphasized usability while highlighting challenges such as setup costs and the need for digital literacy. The solar-powered design ensures reliable offline operation, consuming roughly 480Wh daily. These solar-powered AI units and fine-tuned models present a sustainable solution to enhance documentation and diagnostics in remote healthcare settings. Real-world trials are crucial to validate system performance, complemented by strategic investments in training, infrastructure, and ethical governance to support scalability. This work has resulted in two provisional patent applications, further advancing its potential for practical deployment.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.001 | 0.003 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.000 | 0.000 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.001 | 0.001 |
| Open science | 0.002 | 0.001 |
| Research integrity | 0.001 | 0.001 |
| Insufficient payload (model declined to judge) | 0.003 | 0.001 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".