Research map / Large language models / AI and society
DIYHealth Suite: Dataset, Model, and Benchmark for Health Management at Home
Changshuo Liu, Wu Junran, Zhongle Xie, Wenqiao Zhang, Kaiping Zheng, Jiaqi Zhu, Qingpeng Cai, Ooi Gene Anne, Marcus Chun Jin Tan, Jianwei Yin, James Wei Luen Yip, Beng Chin Ooi
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| add Shared Low-Rank Expert Mixture and Hyper LoRA (full H2LoRA) over no H2LoRA (ablation) | DIYHealthGPT · closed-QA avg (S2D, Heart, Food) | ACC: 84.83 -> 92.90 | Improved |
| add H2LoRA (ablation) | DIYHealthGPT · closed-QA avg (S2D, Heart, Food) | MCC: 79.79 -> 90.56 | Improved |
| add H2LoRA (ablation) | DIYHealthGPT · open-QA avg (S2D, Heart, Food) | F1-BioBERT: 87.52 -> 88.88 | Improved |
| add H2LoRA (ablation) | DIYHealthGPT · open-QA avg (S2D, Heart, Food) | ROUGE-L: 51.41 -> 56.23 | Improved |
| add Expert Mixture to Hyper LoRA (ablation) | DIYHealthGPT · closed-QA avg (S2D, Heart, Food) | ACC: 91.54 -> 92.90 | Improved |
| add Hyper LoRA to Expert Mixture (ablation) | DIYHealthGPT · closed-QA avg (S2D, Heart, Food) | ACC: 91.56 -> 92.90 | Improved |
| add Hyper LoRA to Expert Mixture (ablation) | DIYHealthGPT · Heart closed-QA | ACC: 79.93 -> 83.10 | Improved |
| add Expert Mixture only (over no H2LoRA) (ablation) | DIYHealthGPT · Heart closed-QA | ACC: 64.79 -> 79.93 | Improved |
| add H2LoRA over LoRA, MoELoRA, HyperLoRA (ablation) | DIYHealthGPT · closed-QA and open-QA | ACC and F1-Bio | Improved |
| add four-stage training progressively (Stage 2 to 4) (ablation) | DIYHealthGPT · DIYHealthBench | ACC and F1-Bio | Improved |
| train DIYHealthGPT instead of fine-tuning Gemma 3-4B or LLaVA-Med on same data | DIYHealthGPT-3.8B · DIYHealthBench | four metrics | Improved |
| DIYHealthGPT vs best baseline Lingshu-7B | DIYHealthGPT-3.8B · closed-QA avg | ACC: 70.74 -> 86.80 | Improved |
Closest papers by meaning
- MedAgentBoard: Benchmarking Multi-Agent Collaboration with Conventional Methods for Diverse Medical Tasks (NeurIPS 2025 · similarity 0.83)
- PhysicianBench: Evaluating LLM Agents in Real-World EHR Environments (NeurIPS 2026 · similarity 0.82)
- MedMemoryBench: Benchmarking Agent Memory in Personalized Healthcare (NeurIPS 2026 · similarity 0.81)
- OpenMHC: Accelerating the Science of Wearable Foundation Models (NeurIPS 2026 · similarity 0.81)
- MDAgents: An Adaptive Collaboration of LLMs for Medical Decision-Making (NeurIPS 2024 · similarity 0.81)
- DearLLM: Enhancing Personalized Healthcare via Large Language Models-Deduced Feature Correlations (AAAI 2025 · similarity 0.81)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
