Pith. sign in

REVIEW 3 major objections 6 minor 2 cited by

Scaling laws in wearable human activity recognition

T0 review · 3 major / 6 minor · reviewed 2026-08-09 · deepseek-v4-flash

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

Pith's one-line read Wearable activity recognition follows power-law scaling with data and model size, and user diversity gives roughly three times steeper gains than more same-user hours.

desk verdict First explicit HAR scaling laws with a user-diversity effect, but the fixed-budget convergence premise needs evidence before the exponents are taken as quantitative guidance. read the letter →

arxiv 2502.03364 v2 pith:NUU4XEOT submitted 2025-02-05 cs.LG

Tom Hoddes , Alex Bijamov , Saket Joshi , Daniel Roggen , Ali Etemad , Robert Harle , David Racz This is my paper · ORCID
classification cs.LG
keywords scalinglawswearablehumanactivityrecognitionself-supervisedpre-trainingmaskedautoencoderpowerlawdatadiversityTransformers
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

This paper aims to establish the first quantitative scaling laws for wearable human activity recognition (HAR). It argues that the pre-training loss of a masked-autoencoder Transformer falls as a power law with both the amount of pre-training data and the number of model parameters, so model capacity can be chosen deliberately rather than by trial and error. The key practical claim is that adding new users to the pre-training set improves the loss about three times faster per hour of data than adding more data from existing users, making data diversity the dominant lever. The paper further shows that these pre-training laws translate into downstream gains on three benchmarks, beating from-scratch training, and that prior HAR studies that fixed small encoders may have underused their datasets.

What carries the argument

The machinery is a masked autoencoder built on a Vision Transformer adapted to raw accelerometer and gyroscope windows: 128 samples at 50Hz, patches of 4 samples, 70% random patch masking, a small 2-block decoder, and a 9-model grid over widths of 512, 1024, and 2048 hidden units and depths of 5, 10, and 20 blocks, spanning 1M to 63M parameters. Pre-training fixes 500,000 steps at batch size 2048, which the paper takes as convergence for all configurations. The scaling behaviour is extracted by fitting power laws $L = aD^{-\alpha}$ for loss versus data hours and $L = bN^{-\beta}$ for loss versus parameters, separately for two sampling strategies, USER, adding whole users, and RANDOM, drawing random examples from all users, and then checking that these fits predict downstream linear-evaluation F1.

What would settle it

Train one encoder architecture on a small and a large data slice from the same Extrasensory fold with step budgets of, say, 125,000, 500,000, and 2 million steps, and check whether the fitted loss-versus-data exponent steepens as the budget grows; if the exponent moves, the reported scaling laws are contaminated by the fixed training budget.

Watch

Extended reading notes

Core claim

The paper's central claim is that HAR pre-training has scaling laws of the same form as language and vision, and that these laws are useful for model design. On the Extrasensory dataset, about 1,589 hours from 60 users, pre-training loss of a masked-autoencoder ViT falls as $L = aD^{-\alpha}$ in data hours and as a power law in parameter count across architectures from 1M to 63M parameters. The fitted exponents are stable across five user-split folds, and the USER sampling exponent, roughly $\alpha \approx 0.044$ to $0.052$, is about three times the RANDOM sampling exponent, roughly $\alpha \approx 0.015$ to $0.020$, so increasing the number of users is the more effective way to scale data. The laws transfer to downstream linear-evaluation F1 on UCI HAR, WISDM Phone, and WISDM Watch, where pre-training outperforms from-scratch training, and the best downstream results come from the largest encoder, suggesting that several earlier HAR studies were under-parameterized.

Load-bearing premise

The paper assumes that every pre-training run, regardless of data size or model size, reaches convergence within a fixed 500,000 training steps, so that loss differences reflect data and capacity rather than unfinished training.

Editorial extensions

If this is right

  • Larger encoders are required to convert additional pre-training data into downstream gains, and the optimal capacity grows monotonically with data size for pre-training loss.
  • Data diversity is a primary lever: spending a fixed data budget on new users gives roughly three times steeper pre-training loss reduction than spending it on additional examples from existing users.
  • Pre-trained HAR models beat from-scratch baselines on UCI HAR, WISDM Phone, and WISDM Watch, with the largest 63M encoder performing best downstream.
  • Previously published HAR studies that fixed small encoder capacities, around 1.5M to 10M parameters, should be revisited with capacities of at least 30M parameters to see whether their reported plateaus disappear.
  • Augmentations such as random rotation and scaling improve downstream performance, especially at larger model scales, and can shift the optimal capacity upward.

Reading between the lines

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

  • Beyond the paper, the fitted exponents imply a practical data-budgeting rule: for a fixed pre-training hour budget, maximizing the number of distinct users should generally beat maximizing hours per user, so collection protocols should prioritize subject coverage over recording length.
  • The paper's fixed 500,000-step convergence assumption is testable: re-fitting the exponents with multiple step budgets would reveal whether large-data runs were under-trained and whether the reported slopes change.
  • Because downstream F1 keeps improving with the largest model even at minimal pre-training data, the loss-based scaling law may understate the value of capacity for classification, and a direct loss-to-F1 transfer model would be a useful follow-up.
  • The same user-versus-random comparison could be applied to other sensor modalities and to contrastive pre-training objectives, which the paper lists as future directions.
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

3 major / 6 minor

Summary. The paper proposes to establish scaling laws for wearable human activity recognition (HAR) by pre-training Masked Autoencoder Vision Transformers on the public Extrasensory dataset. The authors perform a grid search over data size (six fractions), sampling strategy (adding users vs. random examples), and model capacity (nine Transformer architectures), with three learning rates each, totaling 1,620 pre-training runs. They report that pre-training test loss follows a power law with respect to data hours and parameter count, that the USER sampling strategy yields an exponent roughly three times larger than RANDOM sampling, and that these trends translate to improved downstream linear-evaluation F1 on UCI HAR and WISDM Phone/Watch. They conclude that model capacity should be increased with data volume and that some previously published HAR works were under-parameterized.

Significance. If the findings hold, the paper would provide a practical, quantitative tool for choosing model capacity and pre-training data diversity in HAR, a domain where such design rules are currently lacking. The diversity result (more users beats more data per user) is potentially important and actionable. Strengths of the work include the use of public datasets, a large and carefully enumerated grid search, attention to the inference-compute constraints typical of HAR, and explicit downstream validation on three benchmarks. The main risks are methodological: the convergence assumption behind the scaling-law extraction is asserted but not demonstrated, the power-law fits are presented without uncertainty quantification, and the downstream trends are based on best-of-many selection without addressing multiple comparisons. These issues bear directly on the central claims, so they must be addressed before the paper can be accepted.

major comments (3)
  1. [Section 3.1, Section 3.6] The claim that 500,000 pre-training steps is 'sufficient for convergence given our model and data sizes' is load-bearing for all of Section 4.2, but no convergence curves, validation curves, or stopping criteria are provided. With a fixed step count and batch size of 2048, smaller data fractions receive many more epochs than larger ones (e.g., 10% data would be over 1000 epochs if 100% is over 100 epochs), so if large-data runs are under-trained, the fitted loss-vs-data exponents in Table 2 and Figure 2 will be systematically flattened. Please provide representative pre-training loss curves for the smallest and largest data fractions and model sizes, for both sampling strategies, and show that the reported losses are within a small tolerance of the converged value, or otherwise qualify the scaling laws as fixed-budget scaling laws rather than converged-loss scaling laws.
  2. [Section 4.2, Table 2, Figure 2] The power-law fits are reported without confidence intervals, goodness-of-fit statistics, residual analysis, or even the number of points used in each fit. With only six data sizes per fold and exponents as small as -0.015 to -0.052, the claim that the USER exponent is 'roughly 3 times greater' than RANDOM needs uncertainty estimates; the observed ratios could easily be within noise. Please report standard errors or bootstrap intervals for the exponents, show the raw per-fold points overlaid with the fitted curves, and describe the fitting procedure (e.g., least squares on log-log loss vs. log hours, and whether intercepts are free per fold).
  3. [Section 4.3, Figures 4 and 5] The downstream performance claims are based on selecting the best F1 score across 27 runs (9 architectures x 3 learning rates) in Figure 4 and across 180 runs in Figure 5. Selecting the best of many runs inflates the expected F1 and can create spurious monotonic trends with data size or capacity, especially since the number of runs per point is not matched across the compared conditions. Please report the mean and standard deviation or the full distribution across the 27 or 180 runs, and show that the best-run trend is consistent with the median-run trend. Without this, the claim that downstream performance 'translates' the pre-training scaling laws is not fully supported.
minor comments (6)
  1. [Section 3.1] The text says 'we do not fix the amount of compute' immediately after stating that pre-training steps are fixed to 500,000 with a fixed batch size; this is confusing and should be clarified (e.g., the authors mean they do not use a compute budget in the Chinchilla sense, but the actual compute per run is fixed by the protocol).
  2. [Section 3.2] The description of the 9 encoder architectures (3 widths x 3 depths) would benefit from a table listing width, depth, and exact parameter count for each model, since parameter count is the independent variable in Figure 3 and Figure 5.
  3. [Abstract and Section 3.5] The abstract mentions 5000 hours of data from 60 users, while Section 3.5 states the filtered dataset is approximately 1589 hours; these numbers should be reconciled to avoid misleading readers about the pre-training data scale.
  4. [Table 1 and Section 4.3] The WISDM Phone F1 of 31.9% from scratch and 34.3% with linear evaluation is very low; the text explains it is due to including all 18 classes, but the caption and text should state this explicitly and cite the class list so readers do not misinterpret the result as a failure of the method.
  5. [Section 4.5, Figure 7] The augmentation comparison uses 'best' scores across data sizes and learning rates; it is unclear whether the same model and data selections were used for the with- and without-augmentation conditions, which could introduce selection bias. Please clarify the selection protocol.
  6. [References] There are several typos and formatting issues, including 'paralellization' in Section 3, 'Téllez Villamizar and other' in the reference list, and inconsistent citation keys in Section 2.3 (e.g., 'Logacjov et al. (2024) Logacjov and Bach (2024)').

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity: the scaling laws are empirical fits to measured losses, and downstream evaluation uses external benchmarks.

full rationale

The paper's central claims are empirical power-law fits to measured pre-training losses (Figures 2 and 3, Table 2), not derivations from model equations, so there is no derivation chain in which a conclusion is reintroduced as an input. Downstream performance (Section 4.3) is measured on external benchmarks (UCI HAR, WISDM Phone, WISDM Watch) and reported as F1 scores, not predicted from the fitted exponents. The USER-vs-RANDOM comparison is an empirical contrast of two sampling protocols at matched total hours; although the 100%-data condition is shared by construction, the intermediate data-size losses are measured, so the steeper USER exponent is an observed empirical relationship rather than a definitional one. The paper's self-citations (Roggen co-authored references in Sections 1 and 2.1) appear only in related-work and baseline context and are not load-bearing for the scaling-law claim. The main validity concern is the unsupported assertion in Sections 3.1 and 3.6 that 500,000 pre-training steps guarantee convergence for all model and data configurations; if false, the fitted exponents could mix optimization-budget effects with data-scaling effects. That is a correctness threat, not circularity, and does not change the circularity score.

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

The central claims rest on modeling choices and empirical assumptions rather than invented entities: power-law functional form, convergence of a fixed training budget, transferability of phone pre-training to watch benchmarks, and the use of MAE loss as a proxy. The main fitted quantities are the per-fold power-law coefficients and the fixed step budget.

free parameters (4)
  • Power-law exponent, USER sampling = -0.044 to -0.052 across folds
    Fitted to Extrasensory pre-training loss vs data hours; the core diversity claim depends on it.
  • Power-law exponent, RANDOM sampling = -0.015 to -0.020 across folds
    Fitted the same way; compared against USER exponent to conclude user diversity yields roughly 3x steeper scaling.
  • Power-law offset per fold = 0.035 to 0.058
    Fitted per fold; offsets differ across test sets according to the authors.
  • Fixed pre-training step budget = 500,000 steps
    Chosen by hand as sufficient for convergence; if wrong, loss-vs-data curves mix data and training-budget effects.
assumptions (4)
  • domain assumption Extrasensory phone IMU data is an appropriate pre-training distribution for HAR and transfers to UCI HAR and WISDM benchmarks.
    Section 3.5 selects Extrasensory for pre-training; Section 4.3 evaluates transfer to UCI/WISDM, including watch data despite phone-only pre-training.
  • domain assumption MAE reconstruction loss is a meaningful proxy for downstream HAR capability.
    Section 4.2 uses pre-training test loss for scaling laws; Section 4.3 then checks downstream F1.
  • ad hoc to paper Power-law form L = a D^-b is the correct functional form for HAR scaling.
    No theoretical derivation is given; the authors fit this form in Figures 2 and 3 and Table 2.
  • ad hoc to paper Fixed 500,000 pre-training steps reaches convergence for all models and data sizes.
    Section 3.1 asserts convergence without showing convergence curves; the scaling-law fits depend on this.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Scaling laws in wearable human activity recognition." pith.science (2026). https://pith.science/paper/NUU4XEOT

@misc{pith2026250203364,
  author       = {Pith},
  title        = {Pith review of: Scaling laws in wearable human activity recognition},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/NUU4XEOT}},
  note         = {Machine review of arXiv:2502.03364}
}
read the original abstract

Many deep architectures and self-supervised pre-training techniques have been proposed for human activity recognition (HAR) from wearable multimodal sensors. Scaling laws have the potential to help move towards more principled design by linking model capacity with pre-training data volume. Yet, scaling laws have not been established for HAR to the same extent as in language and vision. By conducting an exhaustive grid search on both amount of pre-training data and Transformer architectures, we establish the first known scaling laws for HAR. We show that pre-training loss scales with a power law relationship to amount of data and parameter count and that increasing the number of users in a dataset results in a steeper improvement in performance than increasing data per user, indicating that diversity of pre-training data is important, which contrasts to some previously reported findings in self-supervised HAR. We show that these scaling laws translate to downstream performance improvements on three HAR benchmark datasets of postures, modes of locomotion and activities of daily living: UCI HAR and WISDM Phone and WISDM Watch. Finally, we suggest some previously published works should be revisited in light of these scaling laws with more adequate model capacities.

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 2 Pith papers

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

  1. Inertia-1: An Open Exploration of Wearable Motion Foundation Models

    cs.LG 2026-07 conditional novelty 6.5 of 10

    Controlled large-scale pretraining on 18.2M hours of wearables shows self-supervised motion models beat scratch training, with triaxial fidelity, data diversity, and task-matched windows mattering more than model size alone.

  2. TinierHAR: Towards Ultra-Lightweight Deep Learning Models for Efficient Human Activity Recognition on Edge Devices

    cs.CV 2025-07 conditional novelty 5.0 of 10

    TinierHAR is an ultra-lightweight HAR model that matches TinyHAR's F1 score with 2.7x fewer parameters and 6.4x fewer MACs across 14 datasets.

Reference graph

Works this paper leans on

42 extracted references · 39 canonical work pages · cited by 2 Pith papers

  1. [1]

    Abbaspourazad, O

    S. Abbaspourazad, O. Elachqar, A. C. Miller, S. Emrani, U. Nallasamy, and I. Shapiro. Large-scale training of foundation models for wearable biosignals. In International Conference on Learning Representations, 2024

    work page 2024

  2. [2]

    Abedin, M

    A. Abedin, M. Ehsanpour, Q. Shi, H. Rezatofighi, and D. C. Ranasinghe. Attend and discriminate: Beyond the state-of-the-art for human activity recognition using wearable sensors. Proceedings of the ACM on Interactive, Mobile, Wearable and Ubiquitous Technologies, 5 0 (1), 2020

    work page 2020

  3. [3]

    M. Bock, A. H\" o lzemann, M. Moeller, and K. Van Laerhoven. Improving deep learning for har with shallow lstms. In ACM International Symposium on Wearable Computers, 2021

    work page 2021

  4. [4]

    S. Chan, Y. Hang, C. Tong, A. Acquah, A. Schonfeldt, J. Gershuny, and A. Doherty. Capture-24: A large dataset of wrist-worn activity tracker data collected in the wild for human activity recognition. Scientific Data, 11 0 (1): 0 1135, 2024

    work page 2024

  5. [5]

    K. Chen, D. Zhang, L. Yao, B. Guo, Z. Yu, and Y. Liu. Deep learning for sensor-based human activity recognition: Overview, challenges, and opportunities. ACM Computing Surveys, 5 0 (4): 0 1--40, 2021

    work page 2021

  6. [6]

    T. Chen, S. Kornblith, M. Norouzi, and G. Hinton. A simple framework for contrastive learning of visual representations. In International Conference on Machine Learning, pages 1597--1607, 2020

    work page 2020

  7. [7]

    Dehghani, J

    M. Dehghani, J. Djolonga, B. Mustafa, P. Padlewski, J. Heek, J. Gilmer, A. P. Steiner, M. Caron, R. Geirhos, I. Alabdulmohsin, R. Jenatton, L. Beyer, M. Tschannen, A. Arnab, X. Wang, C. Riquelme Ruiz, M. Minderer, J. Puigcerver, U. Evci, M. Kumar, S. V. Steenkiste, G. F. Elsayed, A. Mahendran, F. Yu, A. Oliver, F. Huot, J. Bastings, M. Collier, A. A. Grit...

    work page 2023

  8. [8]

    Demrozi, G

    F. Demrozi, G. Pravadelli, A. Bihorac, and P. Rashidi. Human activity recognition using inertial, physiological and environmental sensors: A comprehensive survey. IEEE Access , 8, 2020

    work page 2020

Show all 42 references
  1. [9]

    S. G. Dhekane, H. Haresamudram, M. Thukral, and T. Pl\" o tz. How much unlabeled data is really needed for effective self-supervised human activity recognition? In ACM International Symposium on Wearable Computers, page 66–70, 2023

    2023

  2. [10]

    Dosovitskiy, L

    A. Dosovitskiy, L. Beyer, A. Kolesnikov, D. Weissenborn, X. Zhai, T. Unterthiner, M. Dehghani, M. Minderer, G. Heigold, S. Gelly, J. Uszkoreit, and N. Houlsby. An image is worth 16x16 words: Transformers for image recognition at scale, 2021

    2021

  3. [11]

    Gu, M.-H

    F. Gu, M.-H. Chung, M. Chignell, S. Valaee, B. Zhou, and X. Liu. A survey on deep learning for human activity recognition. ACM Computing Surveys, 54 0 (8), 2021

    2021

  4. [12]

    Haresamudram, A

    H. Haresamudram, A. Beedu, V. Agrawal, P. L. Grady, I. Essa, J. Hoffman, and T. Pl\" o tz. Masked reconstruction based self-supervision for human activity recognition. In ACM International Symposium on Wearable Computers, page 45–49, 2020

    2020

  5. [13]

    Haresamudram, I

    H. Haresamudram, I. Essa, and T. Pl\" o tz. Assessing the state of self-supervised human activity recognition using wearables. Proceedings of the ACM on Interactive, Mobile, Wearable and Ubiquitous Technologies, 6 0 (3), 2022

    2022

  6. [14]

    K. He, X. Chen, S. Xie, Y. Li, P. Doll \'a r, and R. Girshick. Masked autoencoders are scalable vision learners. In IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 16000--16009, 2022

    2022

  7. [15]

    Hoffmann, S

    J. Hoffmann, S. Borgeaud, A. Mensch, E. Buchatskaya, T. Cai, E. Rutherford, D. de Las Casas, L. A. Hendricks, J. Welbl, A. Clark, T. Hennigan, E. Noland, K. Millican, G. van den Driessche, B. Damoc, A. Guy, S. Osindero, K. Simonyan, E. Elsen, J. W. Rae, O. Vinyals, and L. Sifr...

    2022

  8. [16]

    Hsu, H.-C

    Y.-L. Hsu, H.-C. Chang, and Y.-J. Chiu. Wearable sport activity classification based on deep convolutional neural network. IEEE Access, 7: 0 170199--170212, 2019

    2019

  9. [17]

    Kaplan, S

    J. Kaplan, S. McCandlish, T. Henighan, T. B. Brown, B. Chess, R. Child, S. Gray, A. Radford, J. Wu, and D. Amodei. Scaling laws for neural language models, 2020

    2020

  10. [18]

    D. Kara, T. Kimura, S. Liu, J. Li, D. Liu, T. Wang, R. Wang, Y. Chen, Y. Hu, and T. Abdelzaher. Freqmae: Frequency-aware masked autoencoder for multi-modal iot sensing. In ACM Web Conference, page 2795–2806, 2024. ISBN 9798400701719

    2024

  11. [19]

    E. Kim, S. Helal, and D. Cook. Human activity recognition and pattern discovery. Pervasive Computing, 9 0 (1): 0 48--53, 2010

    2010

  12. [20]

    Lee and B

    S. Lee and B. Eskofier. Special issue on wearable computing and machine learning for applications in sports, health, and medical engineering. Applied Sciences, 8 0 (167), 2018

    2018

  13. [21]

    Logacjov

    A. Logacjov. Self-supervised learning for accelerometer-based human activity recognition: A survey. Proceedings of the ACM on Interactive, Mobile, Wearable and Ubiquitous Technologies, 8 0 (4), 2024

    2024

  14. [22]

    Logacjov and K

    A. Logacjov and K. Bach. Self-supervised learning with randomized cross-sensor masked reconstruction for human activity recognition. Engineering Applications of Artificial Intelligence, 128: 0 107478, 2024. ISSN 0952-1976

    2024

  15. [23]

    Logacjov, S

    A. Logacjov, S. Herland, A. Ustad, and K. Bach. Selfpab: large-scale pre-training on accelerometer data for human activity recognition. Applied Intelligence, 54 0 (6): 0 4545--4563, 2024

    2024

  16. [24]

    Lukowicz, O

    P. Lukowicz, O. Amft, D. Roggen, and J. Cheng. On-body sensing: From gesture-based input to activity-driven interaction. IEEE Computer, 43 0 (10): 0 92--96, 2010

    2010

  17. [25]

    unzner, P. Schmidt, A. Reiss, M. Hanselmann, R. Stiefelhagen, and R. D\

    S. M\"unzner, P. Schmidt, A. Reiss, M. Hanselmann, R. Stiefelhagen, and R. D\" u richen. Cnn-based sensor fusion techniques for multimodal human activity recognition. In ACM International Symposium on Wearable Computers, 2017

    2017

  18. [26]

    Nakkiran, G

    P. Nakkiran, G. Kaplun, Y. Bansal, T. Yang, B. Barak, and I. Sutskever. Deep double descent: Where bigger models and more data hurt. Journal of Statistical Mechanics: Theory and Experiment, 2021 0 (12): 0 124003, 2021

    2021

  19. [27]

    Narayanswamy, X

    G. Narayanswamy, X. Liu, K. Ayush, Y. Yang, X. Xu, S. Liao, J. Garrison, S. Tailor, J. Sunshine, Y. Liu, T. Althoff, S. Narayanan, P. Kohli, J. Zhan, M. Malhotra, S. Patel, S. Abdel-Ghaffar, and D. McDuff. Scaling wearable foundation models, 2024

    2024

  20. [28]

    Nguyen, C

    D.-A. Nguyen, C. Pham, and N.-A. Le-Khac. Virtual fusion with contrastive learning for single sensor-based activity recognition. IEEE Sensors Journal, 2024

    2024

  21. [29]

    F. J. Ord\' o \ n ez Morales and D. Roggen. Deep convolutional and lstm recurrent neural networks for multimodal wearable activity recognition. Sensors, 16 0 (1): 0 1--25, 2016

    2016

  22. [30]

    Pellatt and D

    L. Pellatt and D. Roggen. Speeding up deep neural architecture search for wearable activity recognition with early prediction of converged performance. Frontiers in Computer Science, 4, 2022

    2022

  23. [31]

    Pl \"o tz and Y

    T. Pl \"o tz and Y. Guan. Deep learning for human activity recognition in mobile computing. Computer, 51 0 (5): 0 50--59, 2018

    2018

  24. [32]

    Reyes-Ortiz, D

    J. Reyes-Ortiz, D. Anguita, A. Ghio, L. Oneto, and X. Parra. Human activity recognition using smartphones. UCI Machine Learning Repository, 2013

    2013

  25. [33]

    San-Segundo, H

    R. San-Segundo, H. Bluck, J. Moreno-Pimentel, A. Stisen, and M. Gil-Martin. Robust human activity recognition using smartwatches and smartphones. Engineering Applications of Artificial Intelligence, 72: 0 190--202, 2018

    2018

  26. [34]

    P. M. Scholl, M. Wille, and K. Van Laerhoven. Wearables in the wetlab: a laboratory system for capturing and guiding experiments. In ACM International Conference on Ubiquitous Computing, 2015

    2015

  27. [35]

    C. E. T\' e llez Villamizar and other. Printed textile-based dry electrodes for impedance plethysmography measurements. In IEEE International Flexible Electronics Technology Conference, 2024

    2024

  28. [36]

    Vaizman, K

    Y. Vaizman, K. Ellis, and G. Lanckriet. Recognizing detailed human context in the wild from smartphones and smartwatches. IEEE Pervasive Computing, 16 0 (4): 0 62--74, 2017

    2017

  29. [37]

    X. Wang, X. Wang, L. T., L. Jin, and M. He. HARNAS : Human activity recognition based on automatic neural architecture search using evolutionary algorithms. Sensors, 21 0 (20), 2021

    2021

  30. [38]

    G. Weiss. Wisdm smartphone and smartwatch activity and biometrics dataset. UCI Machine Learning Repository, 2019

    2019

  31. [39]

    Welbourne and E

    E. Welbourne and E. Munguia Tapia. Crowdsignals: A call to crowdfund the community’s largest mobile dataset. In Adjunct Proceedings of Ubicomp, 2014

    2014

  32. [40]

    H. Yuan, S. Chan, A. P. Creagh, C. Tong, A. Acquah, D. A. Clifton, and A. Doherty. Self-supervised learning for human activity recognition using 700,000 person-days of wearable data. NPJ Digital Medicine, 7 0 (1): 0 91, 2024

    2024

  33. [41]

    X. Zhai, A. Kolesnikov, N. Houlsby, and L. Beyer. Scaling vision transformers. In IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 12104--12113, 2022

    2022

  34. [42]

    Y. Zhou, H. Zhao, Y. Huang, T. Riedel, M. Hefenbrock, and M. Beigl. TinyHAR : A lightweight deep learning model designed for human activity recognition. In ACM International Symposium on Wearable Computers, 2022

    2022

Pith tools

Reviewed August 9, 2026 · model on record in the stance chip above.