Towards Learning Discrete Representations via Self-Supervision for Wearables-Based Human Activity Recognition
Open accessThe DOI matches its Crossref record; 48 of the 81 references it cites were found in a registry or an index; 11 of the 32 sources its claims cite were checked for support.
Abstract
Human activity recognition (HAR) in wearable and ubiquitous computing typically involves translating sensor readings into feature representations, either derived through dedicated pre-processing procedures or integrated into end-to-end learning approaches. Independent of their origin, for the vast majority of contemporary HAR methods and applications, those feature representations are typically continuous in nature. That has not always been the case. In the early days of HAR, discretization approaches had been explored—primarily motivated by the desire to minimize computational requirements on HAR, but also with a view on applications beyond mere activity classification, such as, for example, activity discovery, fingerprinting, or large-scale search. Those traditional discretization approaches, however, suffer from substantial loss in precision and resolution in the resulting data representations with detrimental effects on downstream analysis tasks. Times have changed, and in this paper, we propose a return to discretized representations. We adopt and apply recent advancements in vector quantization (VQ) to wearables applications, which enables us to directly learn a mapping between short spans of sensor data and a codebook of vectors, where the index comprises the discrete representation, resulting in recognition performance that is at least on par with their contemporary, continuous counterparts—often surpassing them. Therefore, this work presents a proof of concept for demonstrating how effective discrete representations can be derived, enabling applications beyond mere activity classification but also opening up the field to advanced tools for the analysis of symbolic sequences, as they are known, for example, from domains such as natural language processing. Based on an extensive experimental evaluation of a suite of wearable-based benchmark HAR tasks, we demonstrate the potential of our learned discretization scheme and discuss how discretized sensor data analysis can lead to substantial changes in HAR.
Overview
CoChat Summary
TL;DR
The paper proposes learning discrete representations for wearable-based human activity recognition by adding vector quantization to an Enhanced CPC self-supervised framework, then classifying the resulting symbolic sequences. Across six benchmarks, the learned discrete representations perform comparably to or better than continuous self-supervised and end-to-end baselines on several tasks, especially locomotion and daily-living activities.
Key findings
- VQ-CPC discrete representations performed best or similar to the best on three benchmark datasets: Mobiact reached a mean of 77.8%, and HHAR improved over Enhanced CPC by around 1.5%.
- Discretization was detrimental for fine-grained gym activities: on Myogym, performance dropped significantly compared with other self-supervised methods, and on MHEALTH it dropped by around 4.8% relative to Enhanced CPC.
- Learned VQ-CPC outperformed traditional SAX and SAX-REPEAT discretization; SAX showed reductions of over 10% for HHAR, Myogym, and Motionsense.
- Adding RoBERTa pre-training on the discrete symbols improved all target datasets: small RoBERTa increased HHAR by 2%, Myogym by 2.8%, Mobiact by 1.6%, Motionsense by 2.5%, and leg-based datasets by around 2.5%; medium RoBERTa increased HHAR by 3% and Myogym by 3.5%.
Methods
- Pre-training used the Capture-24 dataset: around 2500 h from 151 participants, single wrist-worn accelerometer, downsampled to 50 Hz; evaluation used six target datasets across wrist, waist, and leg locations, with two datasets per location.
- The framework combines Enhanced CPC self-supervision with online K-means vector quantization; base setup used G=2 groups and V=100 codebook variables, with a GRU/LSTM classifier and MLP backend, evaluated via five-fold cross-validation and five randomized runs.
- Additional experiments varied learned alphabet size, encoder output frequency, base self-supervised method, and RoBERTa small/medium pre-training on discrete sequences.
Limitations
- The paper states this is a proof of concept and that discretization is detrimental for fine-grained movements, as seen on Myogym; it also notes that the medium RoBERTa model did not improve over the small one, likely because Capture-24 is not large enough to leverage the bigger architecture.
- The paper states that lower encoder output frequency substantially reduces performance, with drops around 10% for HHAR, over 15% for Myogym/MHEALTH/PAMAP2, and over 20% for waist-based datasets.
- Evident from the text: pre-training and evaluation use a single accelerometer, and evaluation covers two datasets per sensor location rather than the three used in the referenced assessment framework.
Model-written summary. Check it against the paper before relying on it.
Integrity signals
Registry
Cites
Statistics
Figures





![Figure 3 Visualizing K-means-based quantization: for each z-vector, the l2 distance is computed to the codebook of vectors ei. The index of the nearest codebook vector comprises the discrete representation, whereas the nearest vector itself is passed as the output (i.e., z^-vector). This figure has been adopted and adapted from [27].](https://figures.cochat.ai/paper-figures/4b7f2202524c4795ad84688139b4ba87/towards-learning-discrete-representations-via-self-supervision-for-wearables-figure-3.jpg)


Verification
Claims the paper makes
9 Valid · 0 Not valid · 13 Not checked · 97 Own results and interpretations
Every source the paper cites for this claim was checked and supports it.
In the paper
“it was shown in [16] that matching the sampling rates between pre-training and fine-tuning is important for optimal performance.”
§ Towards Learning Discrete Representations via Self-Supervision for Wearables-Based Human Activity Recognition > 4. Setup > 4.2. Data Pre-Processing
What the sources say
- Supportedcites [16] Haresamudram H. Essa I. Plötz T. Assessing the State of Self-Supervised Human Activity Recognition Using Wearables Proc. ACM Interact. Mob. Wearable Ubiquitous Technol. 2022 6 116 10.1145/3550299 source
Section 5.1.3, “Variations in the Sampling Rate,” compares downstream activity-recognition performance with and without matching the target datasets’ sampling rates to the pre-training dataset. It reports reduced performance when rates differ and concludes that matching them is preferable. (LLM, conf=0.94)
plainstates a cause- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Supported
Every source the paper cites for this claim was checked and supports it.
In the paper
“we only pre-train on randomly sampled 10% of the windows from the train split, as it was shown to have comparable performance to using the entire split in [16]”
§ Towards Learning Discrete Representations via Self-Supervision for Wearables-Based Human Activity Recognition > 4. Setup > 4.2. Data Pre-Processing
What the sources say
- Supportedcites [16] Haresamudram H. Essa I. Plötz T. Assessing the State of Self-Supervised Human Activity Recognition Using Wearables Proc. ACM Interact. Mob. Wearable Ubiquitous Technol. 2022 6 116 10.1145/3550299 source
The 'Effect of the number of training windows on activity recognition' section states that windows were randomly sub-sampled from the Capture-24 train split at {0.01, 0.1, 1, 10, 50, 100}% and concludes, as a general trend, that self-supervised methods 'obtain comparable if not better performance when utilizing only 10% (or ~400k) of training windows'. §5.2.1 likewise notes 'excellent performance can be obtained while using 10% of the training windows', and the 'Source dataset imbalance' section explicitly refers back to Fig. 4 as having 'shown ... comparable performance to using the entire train split'. The claim's qualifiers (random sampling, 10% of windows, comparison to full split) are all present in our copy. (LLM, conf=0.90)
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Supported
Every source the paper cites for this claim was checked and supports it.
In the paper
“We extend this analysis by applying the encoder from multi-task self-supervision. Ref. [15] instead of the base encoder, for pre-training. As the encoder contains three blocks with filter sizes of (24, 16, 8), a total of 46 timesteps (i.e., 0.92 s of movements) contribute to each symbol. We observe a significant drop in performance as a result, with around 15% reduction for HHAR and approx. 30% De”
§ Towards Learning Discrete Representations via Self-Supervision for Wearables-Based Human Activity Recognition > 6. Discussion > 6.2. Analyzing the VQ-CPC Discretization Framework
What the sources say
- Supportedcites [15] Saeed A. Ozcelebi T. Lukkien J. Multi-task self-supervised learning for human activity detection Proc. Acm Interact. Mobile Wearable Ubiquitous Technol. 2019 3 1 30 10.1145/3328932 source
Split into 2 atoms — 2 confirmed, 0 partially supported, 0 disputed, 0 misattributed, 0 unverifiable. Expand the claim to see per-atom verdicts.
plainstates a cause- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Supported
The cited sources support only part of this claim — read each source’s verdict below.
In the paper
“The established multi-task self-sup. Ref. [15] framework is ill-suited for discrete representation learning, with significant reductions in performance throughout, consistently performing worse by approx. 10% for most datasets and peaking at around 29% for Mobiact.”
§ Towards Learning Discrete Representations via Self-Supervision for Wearables-Based Human Activity Recognition > 6. Discussion > 6.2. Analyzing the VQ-CPC Discretization Framework
What the sources say
- Partly supportedcites [15] Saeed A. Ozcelebi T. Lukkien J. Multi-task self-supervised learning for human activity detection Proc. Acm Interact. Mobile Wearable Ubiquitous Technol. 2019 3 1 30 10.1145/3328932 sourceExperimental — accuracy not yet measured
Our copy confirms the identification half of the sentence: Saeed et al. (2019) is the multi-task self-supervised framework for HAR ('we explore multi-task self-supervision for learning representations from sensory data through utilizing transformations of the signals', 5 RELATED WORK, p. 22; 3 APPROACH, p. 6). The rest of what is attributed is not in it: no passage addresses discrete representation learning, quantization or codebooks at all, and nothing reports the ~10% deficit on most datasets or the ~29% peak on MobiAct — the paper's own MobiAct numbers (Table, p. 14 and Appendix Table 4, p. 28) show its self-supervised features at or above the supervised baseline (e.g. kappa 0.8883 vs 0.876) and Fig. 5 (p. 14) states multi-task learning improves performance in all cases. The nearest the source comes to the discrete/continuous contrast is its self-description as 'a suitable candidate for continuous unsupervised representation learning in-the-wild' (3.2 Self-supervised Tasks, p. 7), which does not establish ill-suitedness for discrete learning or the stated magnitudes. (LLM, conf=0.72)
plainstates a cause- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Partly supported
Every source the paper cites for this claim was checked and supports it.
In the paper
“The established multi-task self-sup. Ref. [15] framework is ill-suited for discrete representation learning, with significant reductions in performance throughout, consistently performing worse by approx. 10% for most datasets and peaking at around 29% for Mobiact.”
§ Towards Learning Discrete Representations via Self-Supervision for Wearables-Based Human Activity Recognition > 6. Discussion > 6.2. Analyzing the VQ-CPC Discretization Framework
What the sources say
- Supportedcites [15] Saeed A. Ozcelebi T. Lukkien J. Multi-task self-supervised learning for human activity detection Proc. Acm Interact. Mobile Wearable Ubiquitous Technol. 2019 3 1 30 10.1145/3328932 source
The citation identifies the multi-task self-supervised framework, which Saeed et al. describe in the Fig. 3 caption (p. 8). The VQ-CPC comparison and percentage reductions are the citing authors’ assessment of that framework, not results attributed to Saeed et al. (LLM, conf=0.87)
boostedstates a cause- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Supported
51 more claims
Create a free account to see all 119 claims and to verify the 21 references still open.
- Probably not a substantial claim. 26
- It may state two separate claims as one. 45
- The quote only partly states this claim. 1
Claims other papers make about this one
Each is checked against this paper’s own text: does it say what the citing paper says it does?
The cited source says what the claim attributes to it.
In the citing paper
“Furthermore, investigations into enhancements of CPC for HAR, as presented by subsequent research [79,82], have yielded fully convolutional architectures that exhibit substantial improvements across diverse datasets and activity types.”
From Past, Present, and Future of Sensor-based Human Activity Recognition Using Wearables: A Surveying Tutorial on a Still Challenging Task (2025) · cites this paper as [79]
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
Each contradiction between two passages listed here was confirmed by an independent check.
Rated minor: it would not touch a result — a value no result rests on, a p-value that stays on the same side of the significance level, or a conflict that is clerical, resolved elsewhere in the paper, or outside its own findings.
In the paper
§ Towards Learning Discrete Representations via Self-Supervision for Wearables-Based Human Activity Recognition > 5. Results > 5.4. Impact of the Encoder’s Output Frequency
“(i) for an output frequency of 50 Hz (i.e., no downsampling relative to the input), we change the stride of the first block to 1 and for the second block, set the kernel size and stride = 1”
§ Towards Learning Discrete Representations via Self-Supervision for Wearables-Based Human Activity Recognition > 3. Methodology > 3.1. Discrete Representation Learning Setup
“The layers consist of (32, 64, 128, 256) channels, respectively, with a kernel size of (4, 1, 1, 1) and a stride of (2, 1, 1, 1). The encoder output frequency is 24.5 Hz, as we obtain 49 z-vectors for each window of 100 timesteps”
Citations
Cites
Language Models are Few-Shot Learners
Brown, T. B., Mann, B., Ryder, N., Subbiah, M., Kaplan, J., Dhariwal, P., Neelakantan, A., Shyam, P., Sastry, G., Askell, A., et al. (2020) · arXiv · doi:10.48550/arxiv.2005.14165 · Cited by 23,372 · cited 2×
“Replacing RoBERTa, larger models such as GPT-2 [67] or GPT-3 [8] can be utilized on larger scale wearable sensor data such as UK Biobank [20], leading to potential classification performance improvements.”
MethodKaugeranna, R., Kaugeranna, E., & Aion, (. S. 4. 6. (2023) · DROPS (Schloss Dagstuhl – Leibniz Center for Informatics) · doi:10.4230/lipics.giscience.2023.43 · Cited by 13,690
A simple framework for contrastive learning of visual representations
Chen, T., Kornblith, S., Norouzi, M., & Hinton, G. (2020) · arXiv · doi:10.48550/arxiv.2002.05709 · Cited by 10,857 · cited 1×
“Siamese contrastive learning using the SimCLR framework [69] was explored in [70].”
MethodBootstrap Your Own Latent: A New Approach to Self-Supervised Learning
Grill, J.-B., Strub, F., Altché, F., Tallec, C., Richemond, P. H., Buchatskaya, E., Doersch, C., Pires, B. A., Guo, Z. D., Azar, M. G., et al. (2020) · arXiv · doi:10.48550/arxiv.2006.07733 · Cited by 3,846 · cited 1×
“Sim-Siam [72] and BYOL [73] also have a Siamese setup, albeit are not trained with contrastive learning.”
BackgroundExploring Simple Siamese Representation Learning
Chen, X., & He, K. (2021) · 2021 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr46437.2021.01549 · Cited by 3,725
TextRank: Bringing Order into Text
Mihalcea, R. F., & Tarau, P. (2004) · Conference on Empirical Methods in Natural Language Processing (EMNLP) · Cited by 3,631
HuBERT: Self-Supervised Speech Representation Learning by Masked Prediction of Hidden Units
Hsu, W.-N., Bolte, B., Tsai, Y.-H. H., Lakhotia, K., Salakhutdinov, R., & Mohamed, A. (2021) · IEEE/ACM Transactions on Audio, Speech, and Language Processing · doi:10.1109/taslp.2021.3122291 · Cited by 2,983
Deep Convolutional and LSTM Recurrent Neural Networks for Multimodal Wearable Activity Recognition
Ordóñez, F., & Roggen, D. (2016) · Sensors · doi:10.3390/s16010115 · Cited by 2,761 · cited 1×
“The de facto approach for obtaining features is to compute them via statistical metrics [6,7] and the empirical cumulative distribution function [8], or to learn them directly from the data itself (e.g., via end-to-end training [9–13] or unsupervised learning [14–16]).”
Methodwav2vec 2.0: A Framework for Self-Supervised Learning of Speech Representations
Baevski, A., Zhou, H., Mohamed, A., & Auli, M. (2020) · arXiv · doi:10.48550/arxiv.2006.11477 · Cited by 2,369
A tutorial on human activity recognition using body-worn inertial sensors
Bulling, A., Blanke, U., & Schiele, B. (2014) · ACM Computing Surveys · doi:10.1145/2499621 · Cited by 1,668 · cited 1×
“The de facto approach for obtaining features is to compute them via statistical metrics [6,7] and the empirical cumulative distribution function [8], or to learn them directly from the data itself (e.g., via end-to-end training [9–13] or unsupervised learning [14–16]).”
MethodExperiencing SAX: A novel symbolic representation of time series
Lin, J., Keogh, E., Wei, L., & Lonardi, S. (2007) · Data Mining and Knowledge Discovery · doi:10.1007/s10618-007-0064-z · Cited by 1,618 · cited 4×
“Discretization of time-series data has traditionally been performed using computational methods such as Symbolic Aggregate approXimation (SAX) [45, 77].”
BackgroundMethodInfluentialIntroducing a New Benchmarked Dataset for Activity Monitoring
Reiss, A., & Stricker, D. (2012) · 2012 16th International Symposium on Wearable Computers · doi:10.1109/iswc.2012.13 · Cited by 1,456
Doherty, A., Jackson, D., Hammerla, N., Plötz, T., Olivier, P., Granat, M. H., White, T., van Hees, V. T., Trenell, M. I., Owen, C. G., et al. (2017) · PLOS ONE · doi:10.1371/journal.pone.0169649 · Cited by 1,250 · cited 3×
“Replacing RoBERTa, larger models such as GPT-2 [95] or GPT-3 [96] or modifications to existing methods (e.g., masking spans of data [27]) can be utilized on larger scale datasets such as UK Biobank [94], leading to potential classification performance improvements.”
BackgroundMethodSelf-citationAn online algorithm for segmenting time series
Keogh, E., Chu, S., Hart, D., & Pazzani, M. (2002) · Proceedings 2001 IEEE International Conference on Data Mining · doi:10.1109/icdm.2001.989531 · Cited by 1,149
(2015) · Cited by 989
Stisen, A., Blunck, H., Bhattacharya, S., Prentow, T. S., Kjærgaard, M. B., Dey, A., Sonne, T., & Jensen, M. M. (2015) · Proceedings of the 13th ACM Conference on Embedded Networked Sensor Systems · doi:10.1145/2809695.2809718 · Cited by 714 · cited 17×
“Further, the reduction in performance of SAX-REPEAT relative to SAX on HHAR, MHEALTH, and PAMAP2, indicates that modifying SAX to apply to multi-variate data is challenging.”
BackgroundInfluentialGenerating diverse high-fidelity images with vq-vae-2
Razavi, A., Oord, A. V. D., & Vinyals, O. (2019) · arXiv · doi:10.48550/arxiv.1906.00446 · Cited by 709 · cited 1×
“The generation of high-fidelity images was shown in [46], which proposed improvements to the autoencoder setup from [29].”
BackgroundYeh, C.-C. M., Zhu, Y., Ulanova, L., Begum, N., Ding, Y., Dau, H. A., Silva, D. F., Mueen, A., & Keogh, E. (2016) · 2016 IEEE 16th International Conference on Data Mining (ICDM) · doi:10.1109/icdm.2016.0179 · Cited by 653 · cited 1×
“Another technique for discovering primitives includes utilizing the matrix profile (and its extensions) [44,45], which also facilitates motif discovery.”
MethodDeep, Convolutional, and Recurrent Models for Human Activity Recognition Using Wearables
Hammerla, N. Y., Halloran, S., & Ploetz, T. (2016) · arXiv · doi:10.48550/arxiv.1604.08880 · Cited by 606 · cited 1×
“In recent years, the use of supervised deep learning [10,11] has resulted in great improvements in performance, relative to more traditional methods involving heuristics [14].”
mHealthDroid: A Novel Framework for Agile Development of Mobile Health Applications
Banos, O., Garcia, R., Holgado-Terriza, J. A., Damas, M., Pomares, H., Rojas, I., Saez, A., & Villalonga, C. (2014) · Lecture Notes in Computer Science · doi:10.1007/978-3-319-13105-4_14 · Cited by 583
SHIMMER™ – A Wireless Sensor Platform for Noninvasive Biomedical Research
Burns, A., Greene, B. R., McGrath, M. J., O'Shea, T. J., Kuris, B., Ayer, S. M., Stroiescu, F., & Cionca, V. (2010) · IEEE Sensors Journal · doi:10.1109/jsen.2010.2045498 · Cited by 576
The BOSS is concerned with time series classification in the presence of noise
Schäfer, P. (2015) · Data Mining and Knowledge Discovery · doi:10.1007/s10618-014-0377-7 · Cited by 560 · cited 1×
“…include: (i) Multivariate Bag-Of-SFA Symbols (MBOSS), [23] which learns symbolic representations but via Symbolic Fourier Approximation (SFA) [39,40]; (ii) Word extraction for time series classification (WEASEL) [41], which is an extension for SFA, and applies the ANOVA f-test to determine…”
MethodAvoiding Overfitting: A Survey on Regularization Methods for Convolutional Neural Networks
Santos, C. F. G. D., & Papa, J. P. (2022) · ACM Computing Surveys · doi:10.1145/3510413 · Cited by 557
Henriksen, A., Haugen Mikalsen, M., Woldaregay, A. Z., Muzny, M., Hartvigsen, G., Hopstock, L. A., & Grimsgaard, S. (2018) · Journal of Medical Internet Research · doi:10.2196/jmir.9157 · Cited by 542 · cited 1×
“The widespread availability of commodity wearables such as smartphones and smart-watches, has resulted in increased interest in their utilization for applications such as sports and fitness tracking [1–5].”
BackgroundEnsembles of Deep LSTM Learners for Activity Recognition using Wearables
Guan, Y., & Plötz, T. (2017) · Proceedings of the ACM on Interactive, Mobile, Wearable and Ubiquitous Technologies · doi:10.1145/3090076 · Cited by 511 · cited 1×
“In recent years, the use of supervised deep learning [10,11] has resulted in great improvements in performance, relative to more traditional methods involving heuristics [14].”
Self-citationMulti-task Self-Supervised Learning for Human Activity Detection
Saeed, A., Ozcelebi, T., & Lukkien, J. (2019) · Proceedings of the ACM on Interactive, Mobile, Wearable and Ubiquitous Technologies · doi:10.1145/3328932 · Cited by 346 · cited 5×
“[15] framework is ill-suited for discrete representation learning, with significant reductions in performance throughout, consistently performing worse by approx.”
BackgroundMethodInfluentialAnalyzing features for activity recognition
Huynh, T., & Schiele, B. (2005) · Proceedings of the 2005 joint conference on Smart objects and ambient intelligence: innovative context-aware services: usages and technologies · doi:10.1145/1107548.1107591 · Cited by 339
Feature Learning for Activity Recognition in Ubiquitous Computing
Plötz, T., Hammerla, N., & Olivier, P. (2011) · Newcastle University · doi:10.25405/ncl.34175319 · Cited by 337
Self-citationUnsupervised speech representation learning using WaveNet autoencoders
Chorowski, J., Weiss, R. J., Bengio, S., & van den Oord, A. (2019) · IEEE/ACM Transactions on Audio, Speech, and Language Processing · doi:10.1109/taslp.2019.2938863 · Cited by 325 · cited 1×
“Unsupervised speech representations were learned using Wavenet autoencoders in [47], which also demonstrated the correspondence between phonemes and the learned symbols.”
MethodBanos, O., Villalonga, C., Garcia, R., Saez, A., Damas, M., Holgado-Terriza, J. A., Lee, S., Pomares, H., & Rojas, I. (2015) · BioMedical Engineering OnLine · doi:10.1186/1475-925x-14-s2-s6 · Cited by 321 · cited 11×
“Further, the reduction in performance of SAX-REPEAT relative to SAX on HHAR, MHEALTH, and PAMAP2, indicates that modifying SAX to apply to multi-variate data is challenging.”
MethodInfluentialA comprehensive survey on regularization strategies in machine learning
Tian, Y., & Zhang, Y. (2022) · Information Fusion · doi:10.1016/j.inffus.2021.11.005 · Cited by 298 · cited 1×
“These methods can sometimes require regularization in order to prevent overfitting and improve generalization [58–60].”
MethodChung, Y.-A., Zhang, Y., Han, W., Chiu, C.-C., Qin, J., Pang, R., & Wu, Y. (2021) · 2021 IEEE Automatic Speech Recognition and Understanding Workshop (ASRU) · doi:10.1109/asru51503.2021.9688253 · Cited by 293
Fast and Accurate Time Series Classification with WEASEL
Schäfer, P., & Leser, U. (2017) · Proceedings of the 2017 ACM on Conference on Information and Knowledge Management · doi:10.1145/3132847.3132980 · Cited by 286 · cited 1×
“…learns symbolic representations but via Symbolic Fourier Approximation (SFA) [39,40]; (ii) Word extraction for time series classification (WEASEL) [41], which is an extension for SFA, and applies the ANOVA f-test to determine the most informative Fourier coefficients and subsequently applies…”
MethodLogical-shapelets: an expressive primitive for time series classification
Mueen, A., Keogh, E., & Young, N. (2011) · Proceedings of the 17th ACM SIGKDD international conference on Knowledge discovery and data mining · doi:10.1145/2020408.2020587 · Cited by 282 · cited 1×
“Similarly, primitives of motion have been derived via shapelets [42] for time series classification [43], wherein each shapelet is a local pattern highly indicative of a class.”
BackgroundReview on Wearable Technology in Sports: Concepts, Challenges and Opportunities
Seçkin, A. Ç., Ateş, B., & Seçkin, M. (2023) · Applied Sciences · doi:10.3390/app131810399 · Cited by 279
Willetts, M., Hollowell, S., Aslett, L., Holmes, C., & Doherty, A. (2018) · Scientific Reports · doi:10.1038/s41598-018-26174-1 · Cited by 250 · cited 1×
“The representations are derived using wrist-based accelerometer data from Capture-24 [31–33]—a large-scale dataset that was collected in-the-wild and, as such, is representative for real-world Ubicomp applications of HAR.”
MethodSAX-VSM: Interpretable Time Series Classification Using SAX and Vector Space Model
Senin, P., & Malinchik, S. (2013) · 2013 IEEE 13th International Conference on Data Mining · doi:10.1109/icdm.2013.52 · Cited by 236 · cited 1×
“Other variations include applying Tf-idf weighting of the SAX features, as explored in SAX-VSM [38].”
MethodProtecting sensory data against sensitive inferences
Malekzadeh, M., Clegg, R. G., Cavallaro, A., & Haddadi, H. (2018) · Spiral (Imperial College London) · Cited by 166
Protecting Sensory Data against Sensitive Inferences
Malekzadeh, M., Clegg, R. G., Cavallaro, A., & Haddadi, H. (2018) · Proceedings of the 1st Workshop on Privacy by Design in Distributed Systems · doi:10.1145/3195258.3195260 · Cited by 154 · cited 10×
“However, for Motionsense, the performance is similar to the best performing model overall, which is Enhanced CPC, once again outperforming other self-supervised and supervised baselines.”
MethodInfluentialContrastive Predictive Coding for Human Activity Recognition
Haresamudram, H., Essa, I., & Plötz, T. (2021) · Proceedings of the ACM on Interactive, Mobile, Wearable and Ubiquitous Technologies · doi:10.1145/3463506 · Cited by 150 · cited 3×
“Contrastive Predictive Coding (CPC) was adopted and applied to wearable sensor data in [68], where future timestep prediction was performed under contrastive learning settings.”
BackgroundMethodInfluentialSelf-citationMasked reconstruction based self-supervision for human activity recognition
Haresamudram, H., Beedu, A., Agrawal, V., Grady, P. L., Essa, I., Hoffman, J., & Plötz, T. (2020) · Proceedings of the 2020 International Symposium on Wearable Computers · doi:10.1145/3410531.3414306 · Cited by 146 · cited 1×
“Transformers were explored for self-supervision in [66], by training to reconstruct only randomly masked timesteps of windows of sensor data from mobile phones.”
Self-citationScalable time series classification
Schäfer, P. (2015) · Data Mining and Knowledge Discovery · doi:10.1007/s10618-015-0441-y · Cited by 142 · cited 1×
“…include: (i) Multivariate Bag-Of-SFA Symbols (MBOSS), [23] which learns symbolic representations but via Symbolic Fourier Approximation (SFA) [39,40]; (ii) Word extraction for time series classification (WEASEL) [41], which is an extension for SFA, and applies the ANOVA f-test to determine the…”
MethodHammerla, N. Y., Kirkham, R., Andras, P., & Ploetz, T. (2013) · Proceedings of the 2013 International Symposium on Wearable Computers · doi:10.1145/2493988.2494353 · Cited by 138
Self-citationDiscovering Characteristic Actions from On-Body Sensor Data
Minnen, D., Starner, T., Essa, I., & Isbell, C. (2006) · 2006 10th IEEE International Symposium on Wearable Computers · doi:10.1109/iswc.2006.286337 · Cited by 132 · cited 5×
“One approach involves deriving ‘characteristic actions’ via motif discovery, as such sequences are statistically unlikely to occur across activities and therefore correspond to important actions within the activity [53].”
BackgroundInfluentialSelf-citationTesting Self-Report Time-Use Diaries against Objective Instruments in Real Time
Gershuny, J., Harms, T., Doherty, A., Thomas, E., Milton, K., Kelly, P., & Foster, C. (2020) · Sociological Methodology · doi:10.1177/0081175019884591 · Cited by 124
ColloSSL: Collaborative Self-Supervised Learning for Human Activity Recognition
Jain, Y., Tang, C. I., Min, C., Kawsar, F., & Mathur, A. (2022) · Proceedings of the ACM on Interactive, Mobile, Wearable and Ubiquitous Technologies · doi:10.1145/3517246 · Cited by 113
Assessing the State of Self-Supervised Human Activity Recognition Using Wearables
Haresamudram, H., Essa, I., & Plötz, T. (2022) · Proceedings of the ACM on Interactive, Mobile, Wearable and Ubiquitous Technologies · doi:10.1145/3550299 · Cited by 107
Self-citationSuh, S., Rey, V. F., & Lukowicz, P. (2023) · Knowledge-Based Systems · doi:10.1016/j.knosys.2022.110143 · Cited by 107
Vector-Quantized Autoregressive Predictive Coding
Chung, Y.-A., Tang, H., & Glass, J. (2020) · Interspeech 2020 · doi:10.21437/interspeech.2020-1228 · Cited by 97
COCOA: Cross Modality Contrastive Learning for Sensor Data
Deldari, S., Xue, H., Saeed, A., Smith, D. V., & Salim, F. D. (2022) · Proceedings of the ACM on Interactive, Mobile, Wearable and Ubiquitous Technologies · doi:10.1145/3550316 · Cited by 92
Human Daily Activity and Fall Recognition Using a Smartphone’s Acceleration Sensor
Chatzaki, C., Pediaditis, M., Vavoulas, G., & Tsiknakis, M. (2017) · Communications in Computer and Information Science · doi:10.1007/978-3-319-62704-5_7 · Cited by 91 · cited 2×
“A Samsung Galaxy S3 smartphone placed freely in a trouser pocket was utilized to four types and twelve locomotion-style + transitionary activities [11] at a rate of 200 Hz.”
MethodOn the role of features in human activity recognition
Haresamudram, H., Anderson, D. V., & Plötz, T. (2019) · Proceedings of the 23rd International Symposium on Wearable Computers · doi:10.1145/3341163.3347727 · Cited by 90 · cited 1×
“Going beyond Restricted Boltzmann Machines (RBMs) and Autoencoders [64,65], recent years have seen the rise of ‘self-supervised learn-ing’, that also utilizes unlabeled data for representation learning.”
BackgroundSelf-citationDiscovering multivariate motifs using subsequence density estimation and greedy mixture learning
Minnen, D. C., Isbell, C. L., Essa, I. A., & Starner, T. E. (2007) · National Conference on Artificial Intelligence · Cited by 82
Self-citationContinual learning in sensor-based human activity recognition: An empirical benchmark analysis
Jha, S., Schiemer, M., Zambonelli, F., & Ye, J. (2021) · Information Sciences · doi:10.1016/j.ins.2021.04.062 · Cited by 80
What Makes Good Contrastive Learning on Small-Scale Wearable-based Tasks?
Qian, H., Tian, T., & Miao, C. (2022) · Proceedings of the 28th ACM SIGKDD Conference on Knowledge Discovery and Data Mining · doi:10.1145/3534678.3539134 · Cited by 67
Detecting leisure activities with dense motif discovery
Berlin, E., & Van Laerhoven, K. (2012) · Proceedings of the 2012 ACM Conference on Ubiquitous Computing · doi:10.1145/2370216.2370257 · Cited by 67 · cited 2×
“…the ANOVA f-test to determine the most informative Fourier coefficients and subsequently applies information gain binning for determining the boundaries; (iii) the Sliding Window and BottomUp (SWAB) algorithm, [21,22] which has been used for detecting leisure activities using dense motif discovery.”
BackgroundMethodExploring Contrastive Learning in Human Activity Recognition for Healthcare
Tang, C. I., Perez-Pozuelo, I., Spathis, D., & Mascolo, C. (2020) · arXiv · doi:10.48550/arxiv.2011.11542 · Cited by 67 · cited 2×
“Siamese contrastive learning using the SimCLR framework [69] was explored in [70].”
MethodGestures are strings: efficient online gesture spotting and classification using string matching
Stiefmeier, T., Roggen, D., & Tröster, G. (2007) · Proceedings of the Second International Conference on Body Area Networks BodyNets · doi:10.4108/bodynets.2007.143 · Cited by 62 · cited 4×
“…complex movements into a smaller, finite alphabet of discrete (symbolic) representations, thereby simplifying tasks such as spotting gestures [17], and recognizing activities [18,19] via the use of efficient algorithms from string matching and bioinformatics, or even simple nearest neighbors…”
BackgroundMethodInfluentialTime Series Classification by Sequence Learning in All-Subsequence Space
Nguyen, T. L., Gsponer, S., & Ifrim, G. (2017) · IEEE International Conference on Data Engineering (ICDE) · doi:10.1109/icde.2017.142 · Cited by 52
Combining optimal clustering and Hidden Markov models for extractive summarization
Fung, P., Ngai, G., & Cheung, C.-S. (2003) · doi:10.3115/1119312.1119315 · Cited by 47
MyoGym: introducing an open gym data set for activity recognition collected using myo armband
Koskimäki, H., Siirtola, P., & Röning, J. (2017) · Proceedings of the 2017 ACM International Joint Conference on Pervasive and Ubiquitous Computing and Proceedings of the 2017 ACM International Symposium on Wearable Computers · doi:10.1145/3123024.3124400 · Cited by 46 · cited 1×
“The widespread availability of commodity wearables such as smartphones and smart-watches, has resulted in increased interest in their utilization for applications such as sports and fitness tracking [1–5].”
BackgroundDeep Auto-Set: A Deep Auto-Encoder-Set Network for Activity Recognition Using Wearables
Varamin, A. A., Abbasnejad, E., Shi, Q., Ranasinghe, D. C., & Rezatofighi, H. (2018) · Proceedings of the 15th EAI International Conference on Mobile and Ubiquitous Systems: Computing, Networking and Services · doi:10.1145/3286978.3287024 · Cited by 44 · cited 1×
“Going beyond Restricted Boltzmann Machines (RBMs) and Autoencoders [64,65], recent years have seen the rise of ‘self-supervised learn-ing’, that also utilizes unlabeled data for representation learning.”
BackgroundZhu, Y., Imamura, M., Nikovski, D. N., & Keogh, E. (2017) · Proceedings · doi:10.1109/icdm.2017.79 · Cited by 43
Pruning Redundant Mappings in Transformer Models via Spectral-Normalized Identity Prior
Lin, Z., Liu, J., Yang, Z., Hua, N., & Roth, D. (2020) · Findings of the Association for Computational Linguistics: EMNLP 2020 · doi:10.18653/v1/2020.findings-emnlp.64 · Cited by 34
Using SAX representation for human action recognition
Junejo, I. N., & Aghbari, Z. A. (2012) · Journal of Visual Communication and Image Representation · doi:10.1016/j.jvcir.2012.05.001 · Cited by 32 · cited 2×
“Such computational discretization methods have also been applied for HAR applications, using SAX [18] and its variants [19].”
MethodInertial-Measurement-Unit-Based Novel Human Activity Recognition Algorithm Using Conformer
Kim, Y.-W., Cho, W.-H., Kim, K.-S., & Lee, S. (2022) · Sensors · doi:10.3390/s22103932 · Cited by 31
Investigating Enhancements to Contrastive Predictive Coding for Human Activity Recognition
Haresamudram, H., Essa, I., & Plötz, T. (2023) · 2023 IEEE International Conference on Pervasive Computing and Communications (PerCom) · doi:10.1109/percom56429.2023.10099197 · Cited by 27 · cited 10×
“In addition, prior works such as [16,30] have utilized it as the base for self-supervised pre-training, allowing us to compare our results against those works.”
MethodResultInfluentialSelf-citationInvestigating Enhancements to Contrastive Predictive Coding for Human Activity Recognition
Haresamudram, H., Essa, I., & Ploetz, T. (2022) · arXiv · Cited by 27 · cited 10×
“In addition, prior works such as [16,30] have utilized it as the base for self-supervised pre-training, allowing us to compare our results against those works.”
MethodResultInfluentialSelf-citationSport Database: Cardiorespiratory data acquired through wearable sensors while practicing sports
Sbrollini, A., Morettini, M., Maranesi, E., Marcantoni, I., Nasim, A., Bevilacqua, R., Riccardi, G. R., & Burattini, L. (2019) · Data in Brief · doi:10.1016/j.dib.2019.104793 · Cited by 26
MBOSS: A Symbolic Representation of Human Activity Recognition Using Mobile Sensors
Montero Quispe, K. G., Sousa Lima, W., Macêdo Batista, D., & Souto, E. (2018) · Sensors · doi:10.3390/s18124354 · Cited by 24 · cited 3×
“Moreover, the discretization methods are more difficult to apply to multichannel sensor data, requiring specialized handling and containing exploding alphabet sizes [57].”
MethodInfluentialWord Segmentation on Discovered Phone Units With Dynamic Programming and Self-Supervised Scoring
Kamper, H. (2023) · IEEE/ACM Transactions on Audio, Speech, and Language Processing · doi:10.1109/taslp.2022.3229264 · Cited by 24 · cited 1×
“Interestingly, the discrete representations enable the unsupervised discovery of acoustic units where phonemes are automatically mapped to a small set of discrete representations, enabling phoneme discovery and segmentation [17, 19, 38, 83].”
BackgroundHuman Activity Recognition Based on Symbolic Representation Algorithms for Inertial Sensors
Lima, W. S., Bragança, H. L. D. S., Quispe, K. G. M., & Souto, E. (2018) · Sensors · doi:10.3390/s18114045 · Cited by 24 · cited 3×
“Many of these methods are surveyed in [78] in order to compare the raw performance as well as other factors such as the time taken for computation etc.”
MethodInfluentialSpatial-Temporal Masked Autoencoder for Multi-Device Wearable Human Activity Recognition
Miao, S., Chen, L., & Hu, R. (2023) · Proceedings of the ACM on Interactive, Mobile, Wearable and Ubiquitous Technologies · doi:10.1145/3631415 · Cited by 23 · cited 1×
“Extending this setup, spatio-temporal masking was explored in [67].”
BackgroundRobust learning from demonstrations using multidimensional SAX
Mohammad, Y. F. O., & Nishida, T. (2014) · International Conference on Control, Automation and Systems (ICCAS) · doi:10.1109/iccas.2014.6987960 · Cited by 20 · cited 2×
“While originally proposed for a single time series, it has been extended to multi-channel data as well, by applying SAX separately to each channel and combining the tuples or first applying Principal Component Analysis (PCA) to reduce to a single channel [37].”
MethodClustering of Human Activities from Wearables by Adopting Nearest Neighbors
Ahmed, A., Haresamudram, H., & Plötz, T. (2022) · ACM International Symposium on Wearable Computers (ISWC) · doi:10.1145/3544794.3558477 · Cited by 14 · cited 1×
“[71] improves clustering performance by leveraging nearest neighbors.”
BackgroundSelf-citationFortes Rey, V., Suh, S., & Lukowicz, P. (2022) · Proceedings of the 2022 ACM International Symposium on Wearable Computers · doi:10.1145/3544794.3558464 · Cited by 12 · cited 1×
“The contrastive setup has also been extended to multiple sensors through approaches such as Learning from the best [75], ColloSSL [76], and CO-COA [77].”
MethodBento, N., Rebelo, J., Carreiro, A. V., Ravache, F., & Barandas, M. (2023) · Sensors · doi:10.3390/s23146511 · Cited by 12 · cited 1×
“It has also been explored for wearable-based HAR, particularly from continual learning [61] and domain generalization [62,63] perspectives.”
BackgroundHow Much Unlabeled Data is Really Needed for Effective Self-Supervised Human Activity Recognition?
Dhekane, S. G., Haresamudram, H., Thukral, M., & Plötz, T. (2023) · Proceedings of the 2023 International Symposium on Wearable Computers · doi:10.1145/3594738.3611366 · Cited by 9 · cited 1×
“[78] examines the pre-training data efficiency, i.e., the minimal quantities of pre-training data required for effective wearable-based self-supervised learning.”
BackgroundSelf-citationUnsupervised document summarization using pre-trained sentence embeddings and graph centrality
Ramirez-Orta, J., & Milios, E. (2021) · Workshop on Scholarly Document Processing (SDP) · doi:10.18653/v1/2021.sdp-1.14 · Cited by 9 · cited 1×
“Approaches include utilized graphs [50, 68] and clustering [22, 24].”
MethodA Temporal Extension of Latent Dirichlet Allocation for Unsupervised Acoustic Unit Discovery
van der Merwe, W., Kamper, H., & Adam du Preez, J. (2022) · Interspeech 2022 · doi:10.21437/interspeech.2022-11369 · Cited by 3
Unsupervised Speech Recognition
Baevski, A., Hsu, W.-N., Conneau, A., & Auli, M. (2021) · arXiv · doi:10.48550/arxiv.2105.11084 · cited 1×
“0 [28], unsupervised speech recognition [50], and VQ-APC [51].”
Background
Research works printed but not linked: 21
- Bock M. Kuehne H. Van Laerhoven K. Moeller M. WEAR: An Outdoor Sports for Wearable and Egocentric Activity Recognition arXiv 2023 2304.05088
- Baevski A. Schneider S. Auli M. vq-wav2vec: Self-supervised learning of discrete speech representations arXiv 2019 1910.05453
- Van Den Oord A. Vinyals O. Neural discrete representation learning Advances in Neural Information Processing Systems MIT Press Cambridge, MA, USA 2017 Volume 30
- Chan Chang S. Doherty A. Capture-24: Activity Tracker Dataset for Human Activity Recognition University of Oxford Oxford, UK 2021
- Liu Y. Ott M. Goyal N. Du J. Joshi M. Chen D. Levy O. Lewis M. Zettlemoyer L. Stoyanov V. Roberta: A robustly optimized bert pretraining approach arXiv 2019 1907.11692
- Devlin J. Chang M.W. Lee K. Toutanova K. Bert: Pre-training of deep bidirectional transformers for language understanding arXiv 2018 1810.04805
- Shieh J. Keogh E. i SAX: Indexing and mining terabyte sized time series Proceedings of the 14th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining Las Legas, NV, USA 24–27 August 2008 623 631
- Ye L. Keogh E. Time series shapelets: A new primitive for data mining Proceedings of the 15th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining Paris, France 28 June–1 July 2009 947 956
- Jang E. Gu S. Poole B. Categorical reparameterization with gumbel-softmax arXiv 2016 1611.01144
- Cuervo S. ańcucki A. Marxer R. Rychlikowski P. Chorowski J. Variable-rate hierarchical CPC leads to acoustic unit discovery in speech arXiv 2022 2206.02211
- Dieleman S. Nash C. Engel J. Simonyan K. Variable-rate discrete representation learning arXiv 2021 2103.06089
- Kukačka J. Golkov V. Cremers D. Regularization for deep learning: A taxonomy arXiv 2017 1710.10686
- Oord A.v.d. Li Y. Vinyals O. Representation learning with contrastive predictive coding arXiv 2018 1807.03748
- Bengio Y. Léonard N. Courville A. Estimating or propagating gradients through stochastic neurons for conditional computation arXiv 2013 1308.3432
- Paszke A. Gross S. Massa F. Lerer A. Bradbury J. Chanan G. Killeen T. Lin Z. Gimelshein N. Antiga L. Pytorch: An imperative style, high-performance deep learning library Advances in Neural Information Processing Systems MIT Press Cambridge, MA, USA 2019 Volume 32
- Kingma D.P. Ba J. Adam: A method for stochastic optimization arXiv 2014 1412.6980
- Lan Z. Chen M. Goodman S. Gimpel K. Sharma P. Soricut R. Albert: A lite bert for self-supervised learning of language representations arXiv 2019 1909.11942
- Sanh V. Debut L. Chaumond J. Wolf T. DistilBERT, a distilled version of BERT: Smaller, faster, cheaper and lighter arXiv 2019 1910.01108
- Vaswani A. Shazeer N. Parmar N. Uszkoreit J. Jones L. Gomez A.N. Kaiser Ł. Polosukhin I. Attention is all you need Advances in Neural Information Processing Systems MIT Press Cambridge, MA, USA 2017 Volume 30
- Radford A. Wu J. Child R. Luan D. Amodei D. Sutskever I. Language models are unsupervised multitask learners Openai Blog 2019 1 9
- Gokhan T. Smith P. Lee M. Extractive financial narrative summarisation using sentencebert based clustering Proceedings of the 3rd Financial Narrative Processing Workshop Lancaster, UK 15–16 September 2021 94 98
Cited by
Haresamudram, H., Tang, C. I., Suh, S., Lukowicz, P., & Plötz, T. (2025) · Proceedings of the ACM on Interactive Mobile Wearable and Ubiquitous Technologies · doi:10.1145/3729467 · Cited by 33
Self-citationThukral, M., Dhekane, S. G., Hiremath, S. K., Haresamudram, H., & Plötz, T. (2025) · Proceedings of the ACM on Interactive Mobile Wearable and Ubiquitous Technologies · doi:10.1145/3712278 · Cited by 52
Self-citationSelf-supervised Learning for Accelerometer-based Human Activity Recognition: A Survey
Logacjov, A. (2024) · Proceedings of the ACM on Interactive Mobile Wearable and Ubiquitous Technologies · doi:10.1145/3699767 · Cited by 40
Thukral, M., Dhekane, S. G., Hiremath, S. K., Haresamudram, H., & Ploetz, T. (2024) · arXiv · doi:10.48550/arxiv.2405.12368 · Cited by 21
Self-citationWang, C., Feng, Y., Zhong, L., Zhu, S.-Y., Zhang, C., Zheng, S.-Q., Liang, C., Wang, Y.-T., He, C.-J., Yu, C., et al. (2023) · Proceedings of the ACM on Interactive Mobile Wearable and Ubiquitous Technologies · doi:10.1145/3643552 · Cited by 13
“Recent advances in action captioning [18, 24, 31, 71, 73] and knowledge-driven natural language generation (NLG) [8, 13, 21, 53] present an opportunity to create a system that can perceive, understand, and offer expert-like feedback on home-based daily functional activities and exercises.”
Background
Metadata
- Authors
- Haresamudram, H., Essa, I., & Plötz, T.
- Affiliations
- Georgia Institute of Technology (US)
- Corresponding author
- Haresamudram, Harish
- Venue
- Sensors
- Publisher
- MDPI AG
- Published
- 2024-02-15
- Volume
- 24
- Issue
- 4
- Pages
- 1238
- ISSN
- 1424-8220 (ISSN-L)
- PMC
- PMC10892435
- OpenAlex
- W4391846608
- Cited by
- 14
- References
- 77
- Citations by year
- 2026: 4 · 2025: 5 · 2024: 4
- FWCI
- 1.82
- Type
- journal article
- Language
- en
- Open access
- gold
- doi.org
- Licence
- Locations
- Sensors (publishedVersion, cc-by)
- PubMed Central (publishedVersion, cc-by)
- PubMed (publishedVersion)
- arXiv (Cornell University) (submittedVersion)
- DOAJ (DOAJ: Directory of Open Access Journals) (submittedVersion)
- arXiv (Cornell University)
- Funders
- NSF (IIS-2112633)
- Topics
- Physical Sciences › Computer Science › Computer Vision and Pattern Recognition › Context-Aware Activity Recognition Systems
- Physical Sciences › Computer Science › Signal Processing › Time Series Analysis and Forecasting
- Physical Sciences › Engineering › Biomedical Engineering › Gait Recognition and Analysis
- Keywords
- human activity recognition, wearable sensors, vector quantization, discrete representations, self-supervised learning, codebook learning, activity classification
- MeSH terms
- Machine Learning, Wearable Electronic Devices, Humans, Human Activities, Natural Language Processing
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
The claim verdicts this report counts are experimental: their accuracy is not yet measured.
A language model drew the claims from the paper on 2026-10-05, under version 3 of the claim definition (model deepseek-v4.1-flash, prompt extract-v3).
A second model labelled each claim on 2026-10-05 (model jev-1.13-20260917, definition revision gold-guideline-v0.2).
Cited sources are checked with a language model’s help (gpt-6-sol and qwen3.8-omni-flash).
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
Data sources
Bibliographic data: Semantic Scholar, OpenAlex, Crossref.
Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).
Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.