Pith. sign in

REVIEW 4 major objections 6 minor 34 references

FemWear: A Specialized Wearable Foundation Model for Women's Health

T0 review · 4 major / 6 minor · reviewed 2026-08-12 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read FemWear repurposes a frozen general wearable encoder with only 1.11% trainable parameters and shows positive transfer for menstrual onset and cramps under strict participant-level validation.

desk verdict A transparent, well-controlled parameter-efficient wearable specialization study whose honest negative results are as valuable as its modest positive transfer; the central claim is plausible but the statistical evidence is thinner than the fixed-split headline numbers suggest. read the letter →

arxiv 2608.08244 v1 pith:ODJF6MQD submitted 2026-08-08 cs.AI

Yifan Wang , Chenzhong Li This is my paper · ORCID
classification cs.AI
keywords wearablefoundationmodelswomen'shealthparameter-efficientadaptationmenstrualmultitasklearninglow-rankadapterscalibrationlongitudinalmodeling
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

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

The reading

FemWear is a parameter-efficient specialization of a general wearable foundation model for women's-health tasks. The paper's central claim is that freezing the pretrained multimodal encoder and training only 239,236 adapter parameters (1.11% of the encoder) yields a shared longitudinal representation that improves selected menstrual and symptom endpoints. On a fixed participant split, cycle-phase macro-F1 rose by 8.15% and cramps, mood-symptom, and sleep-problem mean absolute error fell by 9.32%, 5.80%, and 9.43%. Under a stricter nested leave-one-participant-out audit across all 42 participants, 24-hour onset, 72-hour onset, and cramps retained positive changes of 2.87%, 6.35%, and 2.19%, while phase, mood, and sleep were flat or negative and no endpoint had a strictly positive corrected confidence interval. The paper proposes this as a reproducible specialized-model formulation and a precise boundary of where transfer occurs, not as proof of universal dominance or clinical validity.

What carries the argument

The load-bearing mechanism is the rank-32 bottleneck adapter inserted into the final two Transformer blocks of a frozen general wearable encoder, so only 239,236 parameters (1.11%) are trainable while the patch projection and the rest of the encoder stay frozen. The adapter update is a zero-initialized residual of the form $\mathrm{Adapter}(H)=H+W_{\mathrm{up}}\,\mathrm{GELU}(W_{\mathrm{down}}\,\mathrm{LN}(H))$, and a three-expert physiological-regime adapter bank computes a soft mixture of similar updates from the pooled daily embedding. On top of the encoder, a gated recurrent unit summarizes the causal daily history, and task-family heads map the participant state to outputs. The menstrual-onset head outputs a three-bin distribution over onset within 24 hours, 24–72 hours, and later, so $P_{24}=p_0$ and $P_{72}=p_0+p_1$ enforce $P_{24}\le P_{72}$ by construction. The whole specialization is trained with a partial-label multitask loss that averages within task domains, so cohorts with more labels do not dominate through task count, and a train-only temperature scaling improves probability calibration.

What would settle it

Running the identical nested leave-one-participant-out audit on a new participant-disjoint menstrual cohort collected under a prespecified protocol would settle the claim: if the bootstrap confidence interval for any of the positive relative changes (24-hour onset, 72-hour onset, cramps) includes zero or is negative, the claimed transfer does not generalize.

Watch

Extended reading notes

Core claim

On the paper's own terms, the discovery is that a frozen general wearable representation can be repurposed into a women's-health-specific representation with a 1.11% trainable-parameter budget, and that the repurposing transfers to a subset of menstrual outcomes rather than across the board. FemWear adds rank-32 residual adapters to the final two Transformer blocks, a causal GRU over daily embeddings, and task-family heads, then trains the specialization on five women's-health cohorts using partial-label multitask learning. The fixed-split gains are substantial for cycle phase (+8.15% macro-F1), cramps (9.32% lower MAE), mood symptoms (5.80% lower MAE), and sleep problems (9.43% lower MAE), but 24-hour onset AUPRC fell by 3.40%. In the stricter 42-participant nested audit, only 24-hour onset, 72-hour onset, and cramps kept positive relative changes, and the paper explicitly reads this as the current evidential boundary: the model enables targeted transfer and coherent probability outputs, but it does not establish universal performance dominance or clinical validity.

Load-bearing premise

The frozen daily cached embeddings must preserve the needed information from raw minute-level sensor signals (temperature trajectory, heart-rate-variability dynamics, movement patterns), because the joint objective never adapts raw features end to end.

Editorial extensions

If this is right

  • A frozen general wearable representation can be specialized for a new health domain with about 1% of the encoder's parameters, making it practical for small research cohorts.
  • Menstrual onset probabilities produced by the three-bin head are logically coherent and can be reliably calibrated with train-only temperature scaling, so they can serve as research-grade outputs even while discrimination stays modest.
  • The evaluation template—participant-cluster bootstrap, Holm correction, capacity-matched baselines, and a nested leave-one-participant-out audit—gives a transparent way to report where transfer is positive, neutral, or negative.
  • Domain-dependent continual specialization shows that adding women's-health data to a frozen adapter does not uniformly help; pregnancy transfer actually worsened, so task-family-aware specialization is necessary.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • If the frozen daily embedding bottleneck is real, then larger gains on menstrual phase and symptoms would likely require adapting at the sub-daily or raw-signal level, or a wearable foundation model with finer temporal resolution.
  • The absence of a few-label advantage suggests that adding independent participants, not simply adding more label days, is the primary path to improving specialization for women's-health tasks.
  • The capacity-matched comparison points to a testable hypothesis: the positive transfer in FemWear may be carried more by the frozen pretrained representation than by the adapter architecture, so replacing the dual-path head with a shared GRU of equal size might preserve most of the benefit.
  • The train-only calibration result is directly reusable: any downstream task using a three-bin nested onset head can import the same temperature-scaling protocol without touching participant-level held-out data.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

4 major / 6 minor

Summary. The paper introduces FemWear, a parameter-efficient specialization of the frozen OpenMHC wearable foundation model for women's-health tasks. FemWear trains 239,236 encoder parameters (1.11% of the encoder) via low-rank residual adapters, a three-expert physiological-regime adapter bank, and causal task-family heads, producing a shared longitudinal representation for menstrual, symptom, affective, sleep/HRV, autonomic, activity, and pregnancy outcomes. The evaluation spans six cohorts with 63 comparable primary metrics, including 33 from women's-health cohorts, and includes a fixed participant split, a nested 42-participant leave-one-participant-out audit, capacity-matched baselines, calibration analysis, missing-history audits, and a label-efficiency audit. The fixed split shows gains for cycle-phase macro-F1, cramps, mood, and sleep, while the nested audit retains positive point estimates for onset and cramps but no endpoint has a strictly positive multiplicity-corrected confidence interval. The paper is carefully hedged and explicitly disclaims universal dominance and clinical validity.

Significance. If the claims hold, FemWear provides a reproducible template for specializing general wearable foundation models to women's-health research with a very small trainable parameter budget. The strengths of the manuscript are real and should be credited: participant-disjoint splits, participant-cluster bootstrap with 2,000 replicates, Holm correction within task families, train-only temperature scaling, capacity-matched baselines under fixed seeds and budgets, and a public code/data-interface release. The honest reporting of negative results, including the failure to outperform shared-GRU and MMoE baselines and the negative label-efficiency result, is a notable strength. However, the central positive transfer claim rests on statistically fragile evidence: the fixed split has only six validation participants, and the nested audit yields no endpoint with a positive corrected confidence interval. The paper is therefore a useful methodological contribution and a clear evidence-boundary statement, but its headline 'targeted transfer' claim should be treated as exploratory rather than established.

major comments (4)
  1. [§5.2, Table 3] The nested leave-one-participant-out audit is the paper's stated evidential boundary, but no endpoint has a strictly positive multiplicity-corrected confidence interval, so the headline 'retained positive changes' (+2.87%, +6.35%, +2.19%) are point estimates that do not survive the paper's own prespecified inference. The abstract and Section 5.2 should state explicitly that the positive-transfer claim is an exploratory trend, not a statistically supported effect; as written, the central claim is not supported by the reported evidence.
  2. [§3.2, §6] Because the OpenMHC encoder is frozen and only daily cached embeddings are adapted, the trained adapters cannot recover any sub-day or raw minute-level structure that the frozen daily representations discard. The paper does not test whether the observed gains reflect menstrual-specific signal rather than reweighting of participant-level covariates (e.g., average temperature or heart-rate level) already present in the frozen embeddings. A concrete control—for example, a linear probe on the same frozen daily embeddings, regressing out participant-level means before adaptation, or an adapter trained on participant-identity-shuffled labels—would be needed to support the 'women's-health specialization' interpretation.
  3. [§5.1, Table 2] The fixed participant split has only six validation participants, so the four endpoints that 'improved in every seed' on that split are not strong evidence of generalization. Since the abstract's first quantitative claims come from this split, the paper should either report the nested audit as the primary result or add an explicit warning in the abstract that the fixed-split gains are development-set results.
  4. [§5.3, Table 4] FemWear failed the preregistered majority-task criteria against both the shared GRU and the MMoE, and no architecture pair produced a Holm-significant advantage. This is honestly reported, but it means the paper does not establish that the proposed dual-path temporal model is superior to equally sized temporal baselines; the architectural contribution should be presented as a negative result or restricted to a post-hoc task-family analysis rather than as a demonstrated advantage.
minor comments (6)
  1. [Figure 3a] The label 'FemMHC' in Figure 3a should be 'FemWear' to match the rest of the text.
  2. [§4.1] The paper reports 69 task outputs and 63 comparable primary metrics; the six excluded outputs should be identified so readers can assess what is not comparable.
  3. [Ethics Statement] The ethics statement says the final manuscript 'must list' source-study ethics approvals and data-use agreements after author verification, which is not a completed disclosure; please add the actual approvals or state clearly that they are pending.
  4. [Table 4] The 'Utility' column in Table 4 is not defined in the text or appendix; please provide the formula or a reference for this metric.
  5. [§5.5] The calibration results report calibrated Brier scores and ECE values but not the uncalibrated baselines; please report both so the 19.8%/30.7% and 88.2%/84.2% reductions are interpretable.
  6. [References] The references for the DEPRESS Fitbit subset, inPHRsym subset, and the wearable HRV/sleep release are missing; the reproducibility checklist itself notes these must be finalized, but a submitted manuscript should include the dataset citations.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: FemWear's transfer claims are evaluated against a frozen external backbone on held-out participants, with all fitted quantities disclosed and limitations acknowledged.

full rationale

The paper's derivation chain is self-contained. FemWear's adapters, GRU, and task heads are trained on training partitions of the six cohorts, and all reported transfer numbers compare FemWear against the frozen OpenMHC backbone using the same downstream protocol, participant splits, seeds, and fixed checkpoints. The nested 42-participant leave-one-participant-out audit evaluates held-out participants, so the +2.87%, +6.35%, and +2.19% changes are not fitted values renamed as predictions. Temperature scaling is explicitly fitted on training participants only, making the calibration ECE reductions genuine held-out measurements. The only definitional constraint is the three-bin onset head, where P24=p0 and P72=p0+p1 enforce P24<=P72 by construction; the paper states this enforcement explicitly in Section 3.3 and reports zero violations only as a sanity check, not as an empirical discovery. Prior work cited for the backbone (OpenMHC), datasets (mcPHASES, pregnancy clock), and methods (adapters, MMoE, GRU) is external to the present authors; there is no load-bearing self-citation, no imported uniqueness theorem, and no ansatz smuggled in via citation. Section 6 explicitly notes that daily cached representations prevent end-to-end adaptation of raw minute-level features, which is a limitation on the mechanism but not a circularity. The manuscript's own limitations—single-seed nested audit, absence of a strictly positive multiplicity-corrected interval, and negative few-label results—are openly stated and weaken the strength of the claims, but they do not make any prediction equivalent to its inputs.

Assumptions & free parameters 7 free parameters · 4 assumptions · 0 invented entities

This is an empirical ML paper rather than a derivation, so there are no invented physical entities. The listed free parameters are modeling choices that affect the results; the axioms are the key domain assumptions the evaluation depends on.

free parameters (7)
  • Adapter rank = 32
    Bottleneck dimension in Eq. (2); chosen by hand and controls specialization capacity.
  • Number of regime experts = 3
    Expert count in the physiological-regime adapter bank (Section 3.2); affects routing and is not derived from data.
  • Sensor-channel dropping probability = 0.35
    Augmentation probability used in continual specialization (Section 3.4).
  • Masked-patch reconstruction ratio = 15%
    Self-supervision objective weight in Section 3.4.
  • Continual pretraining step budget = 1000
    Fixed number of optimization steps; final checkpoint used (Section 3.4).
  • Learning rate = 1e-4
    AdamW learning rate for continual pretraining (Section 4.4).
  • Temperature scaling scalar = one scalar per task
    Fitted on training participants only for calibration (Section 3.3); reduces ECE but is a fitted parameter.
assumptions (4)
  • domain assumption The OpenMHC LSM2 frozen encoder provides a sufficient representation substrate for women's-health tasks.
    Section 3.2 reuses the pretrained projection and Transformer without fine-tuning; if the frozen daily embeddings lack task-relevant signal, the adapters cannot recover it, and the paper acknowledges this in Section 6.
  • domain assumption mcPHASES participant labels (symptom diaries, urine hormone measurements) are accurate enough to serve as ground truth.
    Table 1 and Section 4.1 depend on the 42-participant mcPHASES cohort; label noise in self-reported symptoms would directly weaken measured transfer.
  • domain assumption Participant-disjoint splits prevent participant leakage into training and evaluation.
    Section 4.1 states participants are disjoint across train, validation, and test; any violation would inflate transfer estimates.
  • domain assumption The single-seed nested leave-one-participant-out audit on 42 participants approximates the target population.
    Section 5.2 uses this audit to define the evidence boundary; low sample size and lack of repeated seeds limit generalizability, as acknowledged in Section 6.

how reviews work

0 comments
Cite this review

Pith. "Pith review of FemWear: A Specialized Wearable Foundation Model for Women's Health." pith.science (2026). https://pith.science/paper/ODJF6MQD

@misc{pith2026260808244,
  author       = {Pith},
  title        = {Pith review of: FemWear: A Specialized Wearable Foundation Model for Women's Health},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/ODJF6MQD}},
  note         = {Machine review of arXiv:2608.08244}
}
read the original abstract

General wearable foundation models are pretrained across broad sensor streams and populations, but are not designed around women's-health tasks. We introduce FemWear, a specialized wearable foundation model that parameter-efficiently repurposes a pretrained multimodal wearable backbone. FemWear retains the patch projection and Transformer encoder, training 239,236 parameters (1.11% of a 21.54M-parameter encoder) through low-rank residual adapters and causal task-family heads. It learns one shared longitudinal representation for menstrual, symptom, affective, sleep/recovery, autonomic, activity, and pregnancy-related outcomes. We evaluate six cohorts with 63 comparable primary metrics, including 33 from women's-health cohorts, while retaining the 32-task OpenMHC ability-retention benchmark. On a fixed participant split over three seeds, FemWear improved cycle-phase macro-F1 by 8.15% and reduced mean absolute error for cramps, mood symptoms, and sleep problems by 9.32%, 5.80%, and 9.43%, respectively. In a stricter 42-participant nested leave-one-participant-out audit, 24-hour onset, 72-hour onset, and cramps retained positive changes of 2.87%, 6.35%, and 2.19%; phase, mood, and sleep were neutral or negative, and no endpoint had a strictly positive corrected confidence interval. Capacity-matched experiments outperformed a latest-day multilayer perceptron but not shared-GRU or multi-gate mixture-of-experts baselines. Train-only calibration reduced onset expected calibration error by 84.2--88.2% with zero temporal-nesting violations. FemWear enables targeted transfer and coherent probability outputs for women's-health research, but does not establish universal performance dominance or clinical validity.

Figures

Figures reproduced from arXiv: 2608.08244 by the authors.

Figure 1
Figure 1. Overview of FemWear. (a) A seven-day multimodal wearable history is formed from activity, heart rate, heart-rate variability, sleep, temperature and oxygen signals. (b) A frozen general wearable foundation encoder maps aligned daily patches to daily embeddings. (c) The trainable FemWear adapter combines causal history with three physiological-regime expert pathways to form a shared female-health state; partial-label… view at source ↗
Figure 2
Figure 2. Menstrual transfer depends on the evaluation protocol. [PITH_FULL_IMAGE:figures/full_fig_p006_2.png] view at source ↗
Figure 3
Figure 3. Capacity control and transfer boundary. (a) Mean final validation loss with standard deviation over seeds 17, 42, and 73 for parameter-matched downstream architectures. Lower is better. (b) Relative domain loss change after women’s-health multi-cohort continual specialization compared with the locked static adapter. Positive values favor continued specialization. and 0.0097. For 72-hour onset, AUPRC and AUROC were 0… view at source ↗
Figures from the paper (1 more)
Figure 4
Figure 4. Figure 4: Reliability analyses. (a) Relative reductions in Brier score and expected calibration error (ECE) after scalar temperature scaling fitted on training participants only. (b) Median task performance under increasingly severe random, contiguous, and latest-day deletions. …

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

34 extracted references · 26 canonical work pages

  1. [2]

    PaPaGei : Open foundation models for optical physiological signals

    Arvind Pillai, Dimitris Spathis, Fahim Kawsar, and Mohammad Malekzadeh. PaPaGei : Open foundation models for optical physiological signals. In International Conference on Learning Representations, 2025. URL https://openreview.net/forum?id=kYwTmlq6Vn

    work page 2025

  2. [3]

    Narayan Schuetz, Yuze Bai, Lianggang Pan, Edgar Eggert, Favour Nerrise, Juan Delgado-SanMartin, Max Rosenblattl, Milana Gurbanova, Mohammad Asadi, Anders Johnson, Paul Schmiedmayer, Dennis Wang, Allan Lawrie, Daniel Seung Kim, Xin Liu, Akshay Paruchuri, Ehsan Adeli, Euan Ashley, and Kelly W. Zhang. OpenMHC : Accelerating the science of wearable foundation...

    work page Pith review arXiv 2026

  3. [4]

    Patel, Khai N

    Gloria Lin, Jin Yi Li, Kelsey Christofferson, Shwetak N. Patel, Khai N. Truong, and Alex Mariakakis. Understanding wrist skin temperature changes to hormone variations across the menstrual cycle. npj Women's Health, 2: 0 35, 2024. doi:10.1038/s44294-024-00037-9

    work page doi:10.1038/s44294-024-00037-9 2024

  4. [5]

    Pulse rate measurement during sleep using wearable sensors, and its correlation with the menstrual cycle phases: A prospective observational study

    Mohaned Shilaih, Vanessa de Clerck, Luca Falco, Frank K \"u bler, and Brigitte Leeners. Pulse rate measurement during sleep using wearable sensors, and its correlation with the menstrual cycle phases: A prospective observational study. Scientific Reports, 7: 0 1294, 2017. doi:10.1038/s41598-017-01433-9

    work page doi:10.1038/s41598-017-01433-9 2017

  5. [7]

    Trevor Standley, Amir Zamir, Dawn Chen, Leonidas Guibas, Jitendra Malik, and Silvio Savarese. Which tasks should be learned together in multi-task learning? In Proceedings of the 37th International Conference on Machine Learning, volume 119 of Proceedings of Machine Learning Research, pages 9120--9132, 2020. URL https://proceedings.mlr.press/v119/standley20a.html

    work page 2020

  6. [8]

    Generalized prompt tuning: Adapting frozen univariate time series foundation models for multivariate healthcare time series

    Mingzhu Liu, Angela Chen, and George Chen. Generalized prompt tuning: Adapting frozen univariate time series foundation models for multivariate healthcare time series. In Proceedings of the 4th Machine Learning for Health Symposium, volume 259 of Proceedings of Machine Learning Research, pages 668--679, 2025. URL https://proceedings.mlr.press/v259/liu25a.html

    work page 2025

  7. [9]

    Repurposing foundation model for generalizable medical time series classification

    Nan Huang, Haishuai Wang, Zihuai He, Marinka Zitnik, and Xiang Zhang. Repurposing foundation model for generalizable medical time series classification. In International Conference on Learning Representations, 2026. URL https://openreview.net/forum?id=wNEzRYiyZM

    work page 2026

  8. [10]

    Ravindra, Camilo Espinosa, Elo \"i se Berson, Thanaphong Phongpreecha, Peinan Zhao, Martin Becker, Alan L

    Neal G. Ravindra, Camilo Espinosa, Elo \"i se Berson, Thanaphong Phongpreecha, Peinan Zhao, Martin Becker, Alan L. Chang, Sayane Shome, Ivana Mari \'c , Davide De Francesco, et al. Deep representation learning identifies associations between physical activity and sleep patterns during pregnancy and prematurity. npj Digital Medicine, 6: 0 171, 2023. doi:10...

    work page doi:10.1038/s41746-023-00911-x 2023

Show all 34 references
  1. [11]

    Parameter-efficient transfer learning for NLP

    Neil Houlsby, Andrei Giurgiu, Stanislaw Jastrzebski, Bruna Morrone, Quentin de Laroussilhe, Andrea Gesmundo, Mona Attariyan, and Sylvain Gelly. Parameter-efficient transfer learning for NLP . Proceedings of Machine Learning Research, 97: 0 2790--2799, 2019. URL https://arxiv.o...

    2019 arXiv

  2. [12]

    Hu, Yelong Shen, Phillip Wallis, Zeyuan Allen-Zhu, Yuanzhi Li, Shean Wang, Lu Wang, and Weizhu Chen

    Edward J. Hu, Yelong Shen, Phillip Wallis, Zeyuan Allen-Zhu, Yuanzhi Li, Shean Wang, Lu Wang, and Weizhu Chen. LoRA : Low-rank adaptation of large language models. In International Conference on Learning Representations, 2022. URL https://openreview.net/forum?id=nZeVKeeFYf9

    2022

  3. [13]

    Jiaqi Ma, Zhe Zhao, Xinyang Yi, Jilin Chen, Lichan Hong, and Ed H. Chi. Modeling task relationships in multi-task learning with multi-gate mixture-of-experts. In Proceedings of the 24th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, pages 1930--193...

    1930

  4. [14]

    Gomez, Lukasz Kaiser, and Illia Polosukhin

    Ashish Vaswani, Noam Shazeer, Niki Parmar, Jakob Uszkoreit, Llion Jones, Aidan N. Gomez, Lukasz Kaiser, and Illia Polosukhin. Attention is all you need. In Advances in Neural Information Processing Systems, volume 30, 2017

    2017

  5. [15]

    Learning phrase representations using RNN encoder--decoder for statistical machine translation

    Kyunghyun Cho, Bart van Merri \"e nboer, Caglar Gulcehre, Dzmitry Bahdanau, Fethi Bougares, Holger Schwenk, and Yoshua Bengio. Learning phrase representations using RNN encoder--decoder for statistical machine translation. In Proceedings of the 2014 Conference on Empirical Met...

    2014 doi

  6. [16]

    Weinberger

    Chuan Guo, Geoff Pleiss, Yu Sun, and Kilian Q. Weinberger. On calibration of modern neural networks. In Proceedings of the 34th International Conference on Machine Learning, volume 70 of Proceedings of Machine Learning Research, pages 1321--1330, 2017. URL https://proceedings....

    2017

  7. [17]

    The precision-recall plot is more informative than the ROC plot when evaluating binary classifiers on imbalanced datasets

    Takaya Saito and Marc Rehmsmeier. The precision-recall plot is more informative than the ROC plot when evaluating binary classifiers on imbalanced datasets. PLOS ONE, 10 0 (3): 0 e0118432, 2015. doi:10.1371/journal.pone.0118432

    2015 doi

  8. [18]

    A simple sequentially rejective multiple test procedure

    Sture Holm. A simple sequentially rejective multiple test procedure. Scandinavian Journal of Statistics, 6 0 (2): 0 65--70, 1979

    1979

  9. [19]

    , journal =

    Schuetz, Narayan and Bai, Yuze and Pan, Lianggang and Eggert, Edgar and Nerrise, Favour and Delgado-SanMartin, Juan and Rosenblattl, Max and Gurbanova, Milana and Asadi, Mohammad and Johnson, Anders and Schmiedmayer, Paul and Wang, Dennis and Lawrie, Allan and Kim, Daniel Seun...

    2026

  10. [20]

    arXiv preprint arXiv:2410.13638 , year =

    Scaling Wearable Foundation Models , author =. arXiv preprint arXiv:2410.13638 , year =

    arXiv

  11. [21]

    2025 , url =

    Pillai, Arvind and Spathis, Dimitris and Kawsar, Fahim and Malekzadeh, Mohammad , booktitle =. 2025 , url =

    2025

  12. [22]

    Proceedings of the 4th Machine Learning for Health Symposium , volume =

    Generalized Prompt Tuning: Adapting Frozen Univariate Time Series Foundation Models for Multivariate Healthcare Time Series , author =. Proceedings of the 4th Machine Learning for Health Symposium , volume =. 2025 , url =

    2025

  13. [23]

    International Conference on Learning Representations , year =

    Repurposing Foundation Model for Generalizable Medical Time Series Classification , author =. International Conference on Learning Representations , year =

  14. [24]

    2025 , note =

    Lin, Blue and Li, Jin Yi and Kalani, Kaavya and Truong, Khai and Mariakakis, Alex , journal =. 2025 , note =. doi:10.13026/zx6a-2c81 , url =

    2025 doi

  15. [25]

    npj Digital Medicine , volume =

    Deep Representation Learning Identifies Associations Between Physical Activity and Sleep Patterns During Pregnancy and Prematurity , author =. npj Digital Medicine , volume =. 2023 , doi =

    2023

  16. [26]

    npj Women's Health , volume =

    Understanding Wrist Skin Temperature Changes to Hormone Variations Across the Menstrual Cycle , author =. npj Women's Health , volume =. 2024 , doi =

    2024

  17. [27]

    Scientific Reports , volume =

    Pulse Rate Measurement During Sleep Using Wearable Sensors, and Its Correlation with the Menstrual Cycle Phases: A Prospective Observational Study , author =. Scientific Reports , volume =. 2017 , doi =

    2017

  18. [28]

    Parameter-Efficient Transfer Learning for

    Houlsby, Neil and Giurgiu, Andrei and Jastrzebski, Stanislaw and Morrone, Bruna and de Laroussilhe, Quentin and Gesmundo, Andrea and Attariyan, Mona and Gelly, Sylvain , journal =. Parameter-Efficient Transfer Learning for. 2019 , url =

    2019

  19. [29]

    and Shen, Yelong and Wallis, Phillip and Allen-Zhu, Zeyuan and Li, Yuanzhi and Wang, Shean and Wang, Lu and Chen, Weizhu , booktitle =

    Hu, Edward J. and Shen, Yelong and Wallis, Phillip and Allen-Zhu, Zeyuan and Li, Yuanzhi and Wang, Shean and Wang, Lu and Chen, Weizhu , booktitle =. 2022 , url =

    2022

  20. [30]

    Proceedings of the 24th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining , pages =

    Modeling Task Relationships in Multi-task Learning with Multi-gate Mixture-of-Experts , author =. Proceedings of the 24th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining , pages =. 2018 , doi =

    2018

  21. [31]

    Proceedings of the 37th International Conference on Machine Learning , pages =

    Which Tasks Should Be Learned Together in Multi-task Learning? , author =. Proceedings of the 37th International Conference on Machine Learning , pages =. 2020 , volume =

    2020

  22. [32]

    Proceedings of the 34th International Conference on Machine Learning , pages =

    On Calibration of Modern Neural Networks , author =. Proceedings of the 34th International Conference on Machine Learning , pages =. 2017 , volume =

    2017

  23. [33]

    Learning Phrase Representations using

    Cho, Kyunghyun and van Merri. Learning Phrase Representations using. Proceedings of the 2014 Conference on Empirical Methods in Natural Language Processing , pages =. 2014 , doi =

    2014

  24. [34]

    Advances in Neural Information Processing Systems , volume =

    Attention Is All You Need , author =. Advances in Neural Information Processing Systems , volume =

  25. [35]

    Scandinavian Journal of Statistics , volume =

    A Simple Sequentially Rejective Multiple Test Procedure , author =. Scandinavian Journal of Statistics , volume =

  26. [36]

    The Precision-Recall Plot Is More Informative than the

    Saito, Takaya and Rehmsmeier, Marc , journal =. The Precision-Recall Plot Is More Informative than the. 2015 , doi =

    2015

Pith tools

Reviewed August 12, 2026 · model on record in the stance chip above.