pith. sign in

arxiv: 2606.07542 · v1 · pith:N2BKWWQ6new · submitted 2026-05-01 · 💻 cs.CY · cs.AI

DIYHealth Suite: Dataset, Model, and Benchmark for Health Management at Home

Changshuo Liu , Junran Wu , Zhongle Xie , Wenqiao Zhang , Kaiping Zheng , Jiaqi Zhu , Qingpeng Cai , Ooi Gene Anne
show 4 more authors
Marcus Chun Jin Tan Jianwei Yin James Wei Luen Yip Beng Chin Ooi
This is my paper

Pith reviewed 2026-07-01 07:37 UTC · model grok-4.3

classification 💻 cs.CY cs.AI
keywords home carehealth managementmultimodal datasetfoundation modellow-rank adaptationbenchmarkgenerative AIDIY health
0
0 comments X

The pith

A 900K multimodal dataset and hybrid adaptation let a new model lead on 11 home health tasks.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper builds DIYHealth Suite to move generative AI health tools from hospitals into homes. It collects DIYHealth-900K, a large set of real home scenarios across modalities, trains DIYHealthGPT with a Hybrid Hyper Low-Rank Adaptation method that adjusts to changing tasks and personal conditions, and releases DIYHealthBench to test models on 11 care tasks. If the approach holds, home users could run diagnosis and management queries with performance that beats both general and medical-specialized models in open and closed settings. Readers would care because portable devices and telemedicine already exist, yet no large shared resources have existed to train or compare systems for this setting.

Core claim

DIYHealthGPT, built on the DIYHealth-900K dataset and driven by Hybrid Hyper Low-Rank Adaptation, reaches state-of-the-art results against general-purpose and medical-specific baselines across 11 home care tasks under both open-QA and closed-QA evaluation on the new DIYHealthBench.

What carries the argument

Hybrid Hyper Low-Rank Adaptation, the technique that lets the model adjust to variable task demands and evolving individual conditions while training on the collected home scenarios.

If this is right

  • Foundation models can now be compared systematically on home care using the released benchmark.
  • The same adaptation approach supports both open-ended and closed-ended queries without separate retraining.
  • Home-collected multimodal data can be turned into usable training resources despite heterogeneity.
  • Performance gains appear across the full spectrum of 11 tasks rather than isolated cases.
  • The framework supplies a concrete starting point for models that must track changing personal health states.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • Portable device makers could feed their sensor streams directly into models trained this way to give real-time guidance.
  • Telemedicine platforms might adopt the benchmark to certify new home-care AI before wider release.
  • Longer-term use could shift routine monitoring away from clinic visits if the model maintains accuracy on longitudinal individual data.
  • The dataset construction steps offer a template for other domains that need large home-collected multimodal collections.

Load-bearing premise

The DIYHealth-900K dataset mirrors the range of actual home care situations and the adaptation method avoids overfitting to the collected examples.

What would settle it

Run DIYHealthGPT on a fresh collection of home data from varied users and devices; if it no longer beats the baselines on the 11 tasks, the central performance claim does not hold.

Figures

Figures reproduced from arXiv: 2606.07542 by Beng Chin Ooi, Changshuo Liu, James Wei Luen Yip, Jianwei Yin, Jiaqi Zhu, Junran Wu, Kaiping Zheng, Marcus Chun Jin Tan, Ooi Gene Anne, Qingpeng Cai, Wenqiao Zhang, Zhongle Xie.

Figure 1
Figure 1. Figure 1: Overview of DIYHealth Suite, integrating DIYHealth-900K, DIYHealthGPT, and DIYHealthBench towards health management at home. for performance evaluation across such diverse applications. To systematically unlock the potential for health manage￾ment at home while addressing the aforementioned key challenges of data accessibility, personalization, and task diversity, we introduce DIYHealth Suite1 , a comprehe… view at source ↗
Figure 2
Figure 2. Figure 2: Illustration of DIYHealth Data Engine (i) Linguistic and Signal Normalizer: We introduce a nor￾malization process that harmonizes heterogeneous raw in￾puts across modalities, including textual symptom descrip￾tions and physiological signals, by resolving medical ab￾breviations, aligning terminology, and standardizing signal preprocessing. This process establishes a shared semantic and statistical basis for… view at source ↗
Figure 4
Figure 4. Figure 4: Model architecture of DIYHealthGPT, where H2LoRA integrates Shared Low-Rank Expert Mixture with Hyper LoRA Adaptation to balance task generalization and instance-level personalization. mechanism, where a global soft-MoE router R dynamically integrates the outputs from the N task-specific H2LoRA blocks. Specifically, the router R assigns mixture weights β conditioned on the shared embedding Z: β = (β 1 , . … view at source ↗
Figure 6
Figure 6. Figure 6: Performance across training stages. S2D DR MAG Diabetes Obesity Heart CKD Food Sleep Skin Oral [PITH_FULL_IMAGE:figures/full_fig_p008_6.png] view at source ↗
Figure 8
Figure 8. Figure 8: Results of clinical expert review [PITH_FULL_IMAGE:figures/full_fig_p008_8.png] view at source ↗
Figure 9
Figure 9. Figure 9: An example prompt to rephrase raw data into natural human conversational expressions for the S2D task in the open-QA setting. MLPs, aligning them with text tokens. This alignment bridges the modality gap between vision and language, and the fused representations are subsequently fed into the LLM to support multimodal understanding and coherent language generation. The number of expert matrices per task K i… view at source ↗
Figure 10
Figure 10. Figure 10: Results of the GPT-5 review. E.3. Hyperparameter Sensitivity Study on Number of Experts We examine the impact of the number of experts on model performance, using MAG, Heart, and Skin tasks, along with the average performance as representative cases. The results are summarized in [PITH_FULL_IMAGE:figures/full_fig_p022_10.png] view at source ↗
Figure 11
Figure 11. Figure 11: Sensitivity study on the number of experts in DIYHealthGPT. strongly agree. For answer-level evaluation, participants are asked to rate the quality of the model-generated answers from DIYHealthGPT across six dimensions related to usability and interaction: • Clarity: This answer is easy to understand. • Usefulness: This answer is useful. • Conciseness: The length is appropriate, not overly long or short. … view at source ↗
Figure 12
Figure 12. Figure 12: Results of user feedback analysis. 23 [PITH_FULL_IMAGE:figures/full_fig_p023_12.png] view at source ↗
Figure 13
Figure 13. Figure 13: Results of inter-rater agreement analysis. F. Discussion and Outlook F.1. Broader Impact Beyond its technical innovations, DIYHealth Suite has the potential to deliver significant impact across age groups and health domains. For children, it can promote dietary education, support oral hygiene tracking, and facilitate developmental health monitoring. For middle-aged individuals, it can assist with stress m… view at source ↗
Figure 14
Figure 14. Figure 14: A case study of the S2D task under closed-QA settings. Some baselines output only the option label, without the corresponding option content. 28 [PITH_FULL_IMAGE:figures/full_fig_p028_14.png] view at source ↗
Figure 15
Figure 15. Figure 15: A case study of the MAG task under closed-QA settings. 29 [PITH_FULL_IMAGE:figures/full_fig_p029_15.png] view at source ↗
Figure 16
Figure 16. Figure 16: A case study of the Diabetes task under closed-QA settings. 30 [PITH_FULL_IMAGE:figures/full_fig_p030_16.png] view at source ↗
Figure 17
Figure 17. Figure 17: A case study of the Obesity task under closed-QA settings. 31 [PITH_FULL_IMAGE:figures/full_fig_p031_17.png] view at source ↗
Figure 18
Figure 18. Figure 18: A case study of the Heart task under closed-QA settings. 32 [PITH_FULL_IMAGE:figures/full_fig_p032_18.png] view at source ↗
Figure 19
Figure 19. Figure 19: A case study of the CKD task under closed-QA settings. 33 [PITH_FULL_IMAGE:figures/full_fig_p033_19.png] view at source ↗
Figure 20
Figure 20. Figure 20: A case study of the Food task under closed-QA settings. 34 [PITH_FULL_IMAGE:figures/full_fig_p034_20.png] view at source ↗
Figure 21
Figure 21. Figure 21: A case study of the Sleep task under closed-QA settings. 35 [PITH_FULL_IMAGE:figures/full_fig_p035_21.png] view at source ↗
Figure 22
Figure 22. Figure 22: A case study of the Skin task under closed-QA settings. 36 [PITH_FULL_IMAGE:figures/full_fig_p036_22.png] view at source ↗
Figure 23
Figure 23. Figure 23: A case study of the Oral task under closed-QA settings. 37 [PITH_FULL_IMAGE:figures/full_fig_p037_23.png] view at source ↗
Figure 24
Figure 24. Figure 24: A case study of the S2D task under open-QA settings. 38 [PITH_FULL_IMAGE:figures/full_fig_p038_24.png] view at source ↗
Figure 25
Figure 25. Figure 25: A case study of the DR task under open-QA settings. 39 [PITH_FULL_IMAGE:figures/full_fig_p039_25.png] view at source ↗
Figure 26
Figure 26. Figure 26: A case study of the MAG task under open-QA settings. 40 [PITH_FULL_IMAGE:figures/full_fig_p040_26.png] view at source ↗
Figure 27
Figure 27. Figure 27: A case study of the Diabetes task under open-QA settings. 41 [PITH_FULL_IMAGE:figures/full_fig_p041_27.png] view at source ↗
Figure 28
Figure 28. Figure 28: A case study of the Obesity task under open-QA settings. 42 [PITH_FULL_IMAGE:figures/full_fig_p042_28.png] view at source ↗
Figure 29
Figure 29. Figure 29: A case study of the Heart task under open-QA settings. 43 [PITH_FULL_IMAGE:figures/full_fig_p043_29.png] view at source ↗
Figure 30
Figure 30. Figure 30: A case study of the CKD task under open-QA settings. 44 [PITH_FULL_IMAGE:figures/full_fig_p044_30.png] view at source ↗
Figure 31
Figure 31. Figure 31: A case study of the Food task under open-QA settings. 45 [PITH_FULL_IMAGE:figures/full_fig_p045_31.png] view at source ↗
Figure 32
Figure 32. Figure 32: A case study of the Sleep task under open-QA settings. 46 [PITH_FULL_IMAGE:figures/full_fig_p046_32.png] view at source ↗
Figure 33
Figure 33. Figure 33: A case study of the Skin task under open-QA settings. 47 [PITH_FULL_IMAGE:figures/full_fig_p047_33.png] view at source ↗
Figure 34
Figure 34. Figure 34: A case study of the Oral task under open-QA settings. 48 [PITH_FULL_IMAGE:figures/full_fig_p048_34.png] view at source ↗
read the original abstract

Generative AI is reshaping healthcare, yet most existing advances rely on hospital-grade devices, which limits their accessibility and potential for health management outside clinical settings. With the proliferation of portable devices and telemedicine, healthcare is shifting toward home-based Diagnosis-It-Yourself (DIY) care. Despite this promise, several distinctive challenges remain: (i) home-collected data are heterogeneous, exacerbated by the absence of standardized large-scale datasets; (ii) models require adaptation to variable task demands and evolving individual conditions; (iii) the broad spectrum of home care tasks lacks a unified benchmark for systematic evaluation. In this paper, we present DIYHealth Suite, a comprehensive framework designed to address these challenges through a tailored dataset, model, and benchmark. We first curate DIYHealth-900K, a large-scale multimodal dataset capturing diverse real-world home care scenarios. Building on this, we propose DIYHealthGPT, an adaptive foundation model for home-based health management, powered by the novel Hybrid Hyper Low-Rank Adaptation technique. Finally, we establish DIYHealthBench, the first benchmark to evaluate foundation models on home care tasks. Extensive experiments demonstrate that DIYHealthGPT delivers state-of-the-art performance over both general-purpose and medical-specific baselines on 11 home care tasks in both open-QA and closed-QA settings, laying the groundwork for the next generation of personalized health management at home.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

2 major / 1 minor

Summary. The paper introduces DIYHealth Suite to address challenges in home-based health management: the DIYHealth-900K multimodal dataset for diverse real-world scenarios, the DIYHealthGPT foundation model using a novel Hybrid Hyper Low-Rank Adaptation technique for task and condition adaptation, and DIYHealthBench as the first unified benchmark for 11 home care tasks. It claims that DIYHealthGPT achieves state-of-the-art performance over general-purpose and medical-specific baselines in both open-QA and closed-QA settings.

Significance. If the experimental claims hold with proper validation, the work would provide valuable resources for accessible, personalized home care AI, filling gaps in heterogeneous data handling, adaptive modeling, and standardized evaluation outside clinical settings.

major comments (2)
  1. [Abstract] Abstract: The central SOTA claim is asserted without any metrics, baselines, data splits, error analysis, or experimental details, rendering it impossible to evaluate whether the dataset, model, or benchmark support the stated performance.
  2. The description of DIYHealth-900K and Hybrid Hyper Low-Rank Adaptation provides no concrete evidence or analysis addressing whether the dataset captures diverse real-world home care scenarios or whether the adaptation avoids overfitting to collected data, which is load-bearing for the adaptation and generalization claims.
minor comments (1)
  1. [Abstract] The abstract uses 'extensive experiments' and 'state-of-the-art' without defining the exact tasks, metrics (e.g., accuracy, F1), or comparison models.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the constructive comments. We agree that the abstract requires more supporting details and that additional concrete evidence and analysis are needed for the dataset diversity and adaptation claims. We will revise accordingly in the next version.

read point-by-point responses

  1. Referee: [Abstract] Abstract: The central SOTA claim is asserted without any metrics, baselines, data splits, error analysis, or experimental details, rendering it impossible to evaluate whether the dataset, model, or benchmark support the stated performance.

    Authors: We agree that the abstract should be more self-contained. In the revised manuscript we will expand the abstract to include key quantitative results (e.g., average improvements over baselines on the 11 tasks), mention the train/validation/test splits, and briefly reference the evaluation protocol and error analysis already present in the experiments section. This will allow readers to assess the SOTA claims without needing to read the full paper first. revision: yes

  2. Referee: [—] The description of DIYHealth-900K and Hybrid Hyper Low-Rank Adaptation provides no concrete evidence or analysis addressing whether the dataset captures diverse real-world home care scenarios or whether the adaptation avoids overfitting to collected data, which is load-bearing for the adaptation and generalization claims.

    Authors: We acknowledge the absence of explicit supporting analysis in the current text. We will add a new subsection that quantifies dataset diversity (e.g., distribution across home environments, device types, patient demographics, and condition severity) and will include targeted experiments for the Hybrid Hyper Low-Rank Adaptation method, such as held-out condition generalization tests, overfitting diagnostics (train vs. validation curves), and ablation comparisons against standard LoRA. These additions will directly address the load-bearing claims. revision: yes

Circularity Check

0 steps flagged

No significant circularity in derivation chain

full rationale

The paper introduces new contributions—a curated multimodal dataset (DIYHealth-900K), a model (DIYHealthGPT) using a novel Hybrid Hyper Low-Rank Adaptation technique, and a benchmark (DIYHealthBench)—then reports empirical SOTA results on 11 tasks. No equations, parameter fits, or self-citations are shown that reduce any central claim to its own inputs by construction. The derivation chain consists of data collection, model training, and external benchmarking, which remains self-contained and falsifiable against independent baselines.

Axiom & Free-Parameter Ledger

0 free parameters · 0 axioms · 0 invented entities

Abstract-only review provides no explicit free parameters, axioms, or invented entities beyond the named contributions; no details on fitting procedures or background assumptions are given.

pith-pipeline@v0.9.1-grok · 5821 in / 1151 out tokens · 25334 ms · 2026-07-01T07:37:18.622365+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Forward citations

Cited by 1 Pith paper

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score.

  1. E-MRL: Cross-view Aligned Evidence-driven Multimodal Reinforcement Learning for Reliable 3D Tumor Analysis

    eess.IV 2026-06 unverdicted novelty 5.0

    E-MRL trains VLMs via RL on a diagnosis-localization-verification MDP with a novel cross-view consistency reward to ground 3D tumor reports in verifiable CT slices.

Reference graph

Works this paper leans on

129 extracted references · 24 canonical work pages · cited by 1 Pith paper · 16 internal anchors

  1. [1]

    Langley , title =

    P. Langley , title =. Proceedings of the 17th International Conference on Machine Learning (ICML 2000) , address =. 2000 , pages =

    2000

  2. [2]

    T. M. Mitchell. The Need for Biases in Learning Generalizations. 1980

    1980

  3. [3]

    M. J. Kearns , title =

  4. [4]

    Machine Learning: An Artificial Intelligence Approach, Vol. I. 1983

    1983

  5. [5]

    R. O. Duda and P. E. Hart and D. G. Stork. Pattern Classification. 2000

    2000

  6. [6]

    Suppressed for Anonymity , author=

  7. [7]

    Newell and P

    A. Newell and P. S. Rosenbloom. Mechanisms of Skill Acquisition and the Law of Practice. Cognitive Skills and Their Acquisition. 1981

    1981

  8. [8]

    A. L. Samuel. Some Studies in Machine Learning Using the Game of Checkers. IBM Journal of Research and Development. 1959

    1959

  9. [9]

    Advances in neural information processing systems , volume=

    Ddxplus: A new dataset for automatic medical diagnosis , author=. Advances in neural information processing systems , volume=

  10. [10]

    Nature Medicine , volume=

    Large language models in medicine , author=. Nature Medicine , volume=. 2023 , publisher=

    2023

  11. [11]

    Nature Machine Intelligence , volume=

    LLM-based agentic systems in medicine and healthcare , author=. Nature Machine Intelligence , volume=. 2024 , publisher=

    2024

  12. [12]

    Advances in Neural Information Processing Systems , volume=

    Language models are few-shot learners , author=. Advances in Neural Information Processing Systems , volume=

  13. [13]

    2023 , url=

    GPT-4 Technical Report , author=. 2023 , url=

    2023

  14. [14]

    Medical image analysis , volume=

    A survey on deep learning in medical image analysis , author=. Medical image analysis , volume=

  15. [15]

    Journal of biomedical informatics , volume=

    Deep EHR: A survey of recent advances on deep learning techniques for electronic health record (EHR) analysis , author=. Journal of biomedical informatics , volume=

  16. [16]

    Briefings in bioinformatics , volume=

    Deep learning for healthcare: review, opportunities and challenges , author=. Briefings in bioinformatics , volume=

  17. [17]

    Advances in neural information processing systems , pages=

    RETAIN: An interpretable predictive model for healthcare using reverse time attention mechanism , author=. Advances in neural information processing systems , pages=

  18. [18]

    npj Digital Medicine , volume=

    Scalable and accurate deep learning with electronic health records , author=. npj Digital Medicine , volume=

  19. [19]

    NPJ digital medicine , volume=

    Deep learning-enabled medical computer vision , author=. NPJ digital medicine , volume=

  20. [20]

    2019 , publisher=

    Deep medicine: how artificial intelligence can make healthcare human again , author=. 2019 , publisher=

    2019

  21. [21]

    IEEE Journal on Emerging and Selected Topics in Circuits and Systems , volume=

    A tinyml platform for on-device continual learning with quantized latent replays , author=. IEEE Journal on Emerging and Selected Topics in Circuits and Systems , volume=. 2021 , publisher=

    2021

  22. [22]

    BMC medical informatics and decision making , volume=

    Explainability for artificial intelligence in healthcare: a multidisciplinary perspective , author=. BMC medical informatics and decision making , volume=. 2020 , publisher=

    2020

  23. [23]

    Advances in Neural Information Processing Systems , volume=

    Visual instruction tuning , author=. Advances in Neural Information Processing Systems , volume=

  24. [24]

    Nature , volume=

    Large language models encode clinical knowledge , author=. Nature , volume=. 2023 , publisher=

    2023

  25. [25]

    The Lancet Digital Health , volume=

    The effect of using a large language model to respond to patient messages , author=. The Lancet Digital Health , volume=. 2024 , publisher=

    2024

  26. [26]

    Proceedings of the Conference on Empirical Methods in Natural Language Processing

    Medclip: Contrastive learning from unpaired medical images and text , author=. Proceedings of the Conference on Empirical Methods in Natural Language Processing. Conference on Empirical Methods in Natural Language Processing , volume=

  27. [27]

    Communications Engineering , volume=

    Interactive computer-aided diagnosis on medical image using large language models , author=. Communications Engineering , volume=. 2024 , publisher=

    2024

  28. [28]

    IEEE Communications magazine , volume=

    A survey of mobile phone sensing , author=. IEEE Communications magazine , volume=. 2010 , publisher=

    2010

  29. [29]

    Pattern recognition letters , volume=

    Deep learning for sensor-based activity recognition: A survey , author=. Pattern recognition letters , volume=. 2019 , publisher=

    2019

  30. [30]

    PLoS Medicine , volume=

    The rise of consumer health wearables: Promises and barriers , author=. PLoS Medicine , volume=. 2016 , publisher=

    2016

  31. [31]

    Cardiology Clinics , volume=

    Using wearable sensors to monitor physical activity in patients with heart failure: a systematic review , author=. Cardiology Clinics , volume=. 2020 , publisher=

    2020

  32. [32]

    Telecommunication Systems , volume=

    A survey on communication components for IoT-based technologies in smart homes , author=. Telecommunication Systems , volume=. 2018 , publisher=

    2018

  33. [33]

    IEEE Communications Surveys & Tutorials , volume=

    Survey in smart grid and smart home security: Issues, challenges and countermeasures , author=. IEEE Communications Surveys & Tutorials , volume=. 2014 , publisher=

    2014

  34. [34]

    IEEE Transactions on Pattern Analysis and Machine Intelligence , volume=

    Self-supervised contrastive representation learning for semi-supervised time-series classification , author=. IEEE Transactions on Pattern Analysis and Machine Intelligence , volume=. 2023 , publisher=

    2023

  35. [35]

    ACM Transactions on Internet of Things , volume=

    A survey of on-device machine learning: An algorithms and learning theory perspective , author=. ACM Transactions on Internet of Things , volume=. 2021 , publisher=

    2021

  36. [36]

    Future Generation Computer Systems , volume=

    Exploiting smart e-Health gateways at the edge of healthcare Internet-of-Things: A fog computing approach , author=. Future Generation Computer Systems , volume=. 2018 , publisher=

    2018

  37. [37]

    Lin, Tianwei and Zhang, Wenqiao and Li, Sijing and Yuan, Yuqian and Yu, Binhe and Li, Haoyuan and He, Wanggui and Jiang, Hao and Li, Mengze and Song, Xiaohui and others , booktitle=

  38. [38]

    Li, Sijing and Lin, Tianwei and Lin, Lingshuai and Zhang, Wenqiao and Liu, Jiang and Yang, Xiaoda and Li, Juncheng and He, Yucheng and Song, Xiaohui and Xiao, Jun and others , booktitle=

  39. [39]

    Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=

    Unified-IO 2: Scaling Autoregressive Multimodal Models with Vision Language Audio and Action , author=. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=

  40. [40]

    Chameleon: Mixed-Modal Early-Fusion Foundation Models

    Chameleon: Mixed-modal early-fusion foundation models , author=. arXiv preprint arXiv:2405.09818 , year=

    work page internal anchor Pith review Pith/arXiv arXiv

  41. [41]

    Journal of Machine Learning Research , volume=

    Palm: Scaling language modeling with pathways , author=. Journal of Machine Learning Research , volume=

  42. [42]

    LLaMA: Open and Efficient Foundation Language Models

    Llama: Open and efficient foundation language models , author=. arXiv preprint arXiv:2302.13971 , year=

    work page internal anchor Pith review Pith/arXiv arXiv

  43. [43]

    International Conference on Learning Representations , year=

    Finetuned Language Models are Zero-Shot Learners , author=. International Conference on Learning Representations , year=

  44. [44]

    Journal of Machine Learning Research , volume=

    Switch transformers: Scaling to trillion parameter models with simple and efficient sparsity , author=. Journal of Machine Learning Research , volume=

  45. [45]

    IEEE Journal of Biomedical and Health Informatics , year=

    Biomedgpt: An open multimodal large language model for biomedicine , author=. IEEE Journal of Biomedical and Health Informatics , year=

  46. [46]

    Proceedings of the 2024 Conference on Empirical Methods in Natural Language Processing , pages=

    Towards injecting medical visual knowledge into multimodal llms at scale , author=. Proceedings of the 2024 Conference on Empirical Methods in Natural Language Processing , pages=

    2024

  47. [47]

    Proceedings of the 23rd workshop on biomedical natural language processing , pages=

    XrayGPT: Chest radiographs summarization using large medical vision-language models , author=. Proceedings of the 23rd workshop on biomedical natural language processing , pages=

  48. [48]

    Advances in Neural Information Processing Systems , volume=

    Llava-med: Training a large language-and-vision assistant for biomedicine in one day , author=. Advances in Neural Information Processing Systems , volume=

  49. [49]

    Cognitive computation , volume=

    Deep learning approach for early detection of Alzheimer’s disease , author=. Cognitive computation , volume=. 2022 , publisher=

    2022

  50. [50]

    International conference on medical image computing and computer-assisted intervention , pages=

    Manifold learning of brain MRIs by deep learning , author=. International conference on medical image computing and computer-assisted intervention , pages=. 2013 , organization=

    2013

  51. [51]

    International conference on medical image computing and computer-assisted intervention , pages=

    Deep feature learning for knee cartilage segmentation using a triplanar convolutional neural network , author=. International conference on medical image computing and computer-assisted intervention , pages=. 2013 , organization=

    2013

  52. [52]

    Machine Learning in Medical Imaging: 5th International Workshop, MLMI 2014, Held in Conjunction with MICCAI 2014, Boston, MA, USA, September 14, 2014

    Deep learning of image features from unlabeled data for multiple sclerosis lesion segmentation , author=. Machine Learning in Medical Imaging: 5th International Workshop, MLMI 2014, Held in Conjunction with MICCAI 2014, Boston, MA, USA, September 14, 2014. Proceedings 5 , pages=. 2014 , organization=

    2014

  53. [53]

    Scientific Reports , volume=

    Computer-aided diagnosis with deep learning architecture: applications to breast lesions in US images and pulmonary nodules in CT scans , author=. Scientific Reports , volume=. 2016 , publisher=

    2016

  54. [54]

    Nature , volume=

    Dermatologist-level classification of skin cancer with deep neural networks , author=. Nature , volume=. 2017 , publisher=

    2017

  55. [55]

    jama , volume=

    Development and validation of a deep learning algorithm for detection of diabetic retinopathy in retinal fundus photographs , author=. jama , volume=. 2016 , publisher=

    2016

  56. [56]

    Proceedings of the 2016 SIAM international conference on data mining , pages=

    Risk prediction with electronic health records: A deep learning approach , author=. Proceedings of the 2016 SIAM international conference on data mining , pages=. 2016 , organization=

    2016

  57. [57]

    Advances in Knowledge Discovery and Data Mining: 20th Pacific-Asia Conference, PAKDD 2016, Auckland, New Zealand, April 19-22, 2016, Proceedings, Part II 20 , pages=

    Deepcare: A deep dynamic memory model for predictive medicine , author=. Advances in Knowledge Discovery and Data Mining: 20th Pacific-Asia Conference, PAKDD 2016, Auckland, New Zealand, April 19-22, 2016, Proceedings, Part II 20 , pages=. 2016 , organization=

    2016

  58. [58]

    Machine Learning for Healthcare Conference , pages=

    Doctor ai: Predicting clinical events via recurrent neural networks , author=. Machine Learning for Healthcare Conference , pages=. 2016 , organization=

    2016

  59. [59]

    Scientific Reports , volume=

    Deep patient: an unsupervised representation to predict the future of patients from the electronic health records , author=. Scientific Reports , volume=. 2016 , publisher=

    2016

  60. [60]

    2014 IEEE International Conference on Bioinformatics and Biomedicine (BIBM) , pages=

    Deep learning for healthcare decision making with EMRs , author=. 2014 IEEE International Conference on Bioinformatics and Biomedicine (BIBM) , pages=. 2014 , organization=

    2014

  61. [61]

    AMIA Summits on Translational Science Proceedings , volume=

    Learning low-dimensional representations of medical concepts , author=. AMIA Summits on Translational Science Proceedings , volume=

  62. [62]

    The Llama 3 Herd of Models

    The llama 3 herd of models , author=. arXiv preprint arXiv:2407.21783 , year=

    work page internal anchor Pith review Pith/arXiv arXiv

  63. [63]

    International conference on machine learning , pages=

    Blip-2: Bootstrapping language-image pre-training with frozen image encoders and large language models , author=. International conference on machine learning , pages=. 2023 , organization=

    2023

  64. [64]

    Machine Learning for Health (ML4H) , pages=

    Med-flamingo: a multimodal medical few-shot learner , author=. Machine Learning for Health (ML4H) , pages=. 2023 , organization=

    2023

  65. [65]

    International Conference on Learning Representations , year=

    LoRA: Low-Rank Adaptation of Large Language Models , author=. International Conference on Learning Representations , year=

  66. [66]

    Nature machine intelligence , volume=

    Parameter-efficient fine-tuning of large-scale pre-trained language models , author=. Nature machine intelligence , volume=. 2023 , publisher=

    2023

  67. [67]

    Diabetes care , volume=

    Globalization of diabetes: the role of diet, lifestyle, and genes , author=. Diabetes care , volume=. 2011 , publisher=

    2011

  68. [68]

    Circulation , volume=

    Dietary and policy priorities for cardiovascular disease, diabetes, and obesity: a comprehensive review , author=. Circulation , volume=. 2016 , publisher=

    2016

  69. [69]

    MoELoRA: Contrastive learning guided mixture of experts on parameter-efficient fine-tuning for large language models,

    Moelora: Contrastive learning guided mixture of experts on parameter-efficient fine-tuning for large language models , author=. arXiv preprint arXiv:2402.12851 , year=

    work page arXiv

  70. [70]

    Proceedings of the IEEE/CVF conference on computer vision and pattern recognition , pages=

    Improved baselines with visual instruction tuning , author=. Proceedings of the IEEE/CVF conference on computer vision and pattern recognition , pages=

  71. [71]

    Proceedings of the 2003 human language technology conference of the North American chapter of the association for computational linguistics , pages=

    Automatic evaluation of summaries using n-gram co-occurrence statistics , author=. Proceedings of the 2003 human language technology conference of the North American chapter of the association for computational linguistics , pages=

    2003

  72. [72]

    Bioinformatics , volume=

    BioBERT: a pre-trained biomedical language representation model for biomedical text mining , author=. Bioinformatics , volume=. 2020 , publisher=

    2020

  73. [73]

    Proceedings of the 40th annual meeting of the Association for Computational Linguistics , pages=

    Bleu: a method for automatic evaluation of machine translation , author=. Proceedings of the 40th annual meeting of the Association for Computational Linguistics , pages=

  74. [74]

    Patterns , volume=

    Evaluating progress in automatic chest x-ray radiology report generation , author=. Patterns , volume=. 2023 , publisher=

    2023

  75. [75]

    International Conference on Learning Representations , year=

    BERTScore: Evaluating Text Generation with BERT , author=. International Conference on Learning Representations , year=

  76. [76]

    HeartcareGPT: A Unified Multimodal ECG Suite for Dual Signal-Image Modeling and Understanding

    Heartcare Suite: Multi-dimensional Understanding of ECG with Raw Multi-lead Signal Modeling , author=. arXiv preprint arXiv:2506.05831 , year=

    work page internal anchor Pith review Pith/arXiv arXiv

  77. [77]

    arXiv preprint arXiv:2306.07971 , year=

    Xraygpt: Chest radiographs summarization using medical vision-language models , author=. arXiv preprint arXiv:2306.07971 , year=

    work page arXiv

  78. [78]

    New England Journal of Medicine , volume=

    The current and future state of AI interpretation of medical images , author=. New England Journal of Medicine , volume=. 2023 , publisher=

    2023

  79. [79]

    Proceedings of the 2023 CHI Conference on Human Factors in Computing Systems , pages=

    The perceived utility of smartphone and wearable sensor data in digital self-tracking technologies for mental health , author=. Proceedings of the 2023 CHI Conference on Human Factors in Computing Systems , pages=

    2023

  80. [80]

    Qwen2.5-VL Technical Report

    Qwen2. 5-vl technical report , author=. arXiv preprint arXiv:2502.13923 , year=

    work page internal anchor Pith review Pith/arXiv arXiv

Showing first 80 references.