Improving Deep Learning for HAR with Shallow LSTMs
Open accessAbstract
Recent studies in Human Activity Recognition (HAR) have shown that Deep Learning methods are able to outperform classical Machine Learning algorithms. One popular Deep Learning architecture in HAR is the DeepConvLSTM. In this paper we propose to alter the DeepConvLSTM architecture to employ a 1-layered instead of a 2-layered LSTM. We validate our architecture change on 5 publicly available HAR datasets by comparing the predictive performance with and without the change employing varying hidden units within the LSTM layer(s). Results show that across all datasets, our architecture consistently improves on the original one: Recognition performance increases up to 11.7% for the F1-score, and our architecture significantly decreases the amount of learnable parameters. This improvement over DeepConvLSTM decreases training time by as much as 48%. Our results stand in contrast to the belief that one needs at least a 2-layered LSTM when dealing with sequential data. Based on our results we argue that said claim might not be applicable to sensor-based HAR.
The claims and statistics of this paper haven’t been checked yet.
Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.
Integrity signals
Registry
Cites
Statistics
Claims the paper makes
The claims and statistics of this paper haven’t been checked yet.
Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.
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?
Using a shallower LSTM in the DeepConvLSTM setup is better for many HAR datasets.
“Bock et al. [16], however, find that using a shallower LSTM in the DeepConvLSTM setup is better for many HAR datasets.”
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 Bock et al. [16]
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
The self-check needs the paper’s full text.
Cites
ImageNet classification with deep convolutional neural networks
Krizhevsky, A., Sutskever, I., & Hinton, G. E. (2017) · Communications of the ACM · doi:10.1145/3065386 · Cited by 112,155 · see citing works
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 10× · see citing works
“Our paper’s contributions are threefold: (1) We show that an altered DeepConvLSTM architecture with a one-layered LSTM overall outperforms architectures employing a two-layered LSTM, by validating our claim using the Opportunity dataset [29] as seen in [26], as well as 4 other popular HAR datasets [28, 32, 33, 36].”
BackgroundMethodInfluentialLearning Hierarchical Features for Scene Labeling
Farabet, C., Couprie, C., Najman, L., & LeCun, Y. (2013) · IEEE Transactions on Pattern Analysis and Machine Intelligence · doi:10.1109/tpami.2012.231 · Cited by 2,620 · see citing works
LSTM-CNN Architecture for Human Activity Recognition
Xia, K., Huang, J., & Wang, H. (2020) · IEEE Access · doi:10.1109/access.2020.2982225 · Cited by 851 · see citing works
Collecting complex activity datasets in highly rich networked sensor environments
Roggen, D., Calatroni, A., Rossi, M., Holleczek, T., Forster, K., Troster, G., Lukowicz, P., Bannach, D., Pirkl, G., Ferscha, A., et al. (2010) · 2010 Seventh International Conference on Networked Sensing Systems (INSS) · doi:10.1109/inss.2010.5573462 · Cited by 788 · see citing works
Deep Learning for Sensor-based Human Activity Recognition
Chen, K., Zhang, D., Yao, L., Guo, B., Yu, Z., & Liu, Y. (2022) · ACM Computing Surveys · doi:10.1145/3447744 · Cited by 690 · cited 2× · see citing works
“[3], based on the findings of Karpathy et al.”
BackgroundMethodInnoHAR: A Deep Neural Network for Complex Human Activity Recognition
Xu, C., Chai, D., He, J., Zhang, X., & Duan, S. (2019) · IEEE Access · doi:10.1109/access.2018.2890675 · Cited by 333 · see citing works
Deep Learning for Sensor-based Human Activity Recognition: Overview, Challenges and Opportunities
Chen, K., Zhang, D., Yao, L., Guo, B., Yu, Z., & Liu, Y. (2020) · arXiv · doi:10.48550/arxiv.2001.07416 · Cited by 314 · cited 2× · see citing works
“[3], based on the findings of Karpathy et al.”
BackgroundMethodPD Disease State Assessment in Naturalistic Environments Using Deep Learning
Hammerla, N., Fisher, J., Andras, P., Rochester, L., Walker, R., & Ploetz, T. (2015) · Proceedings of the AAAI Conference on Artificial Intelligence · doi:10.1609/aaai.v29i1.9484 · Cited by 179 · see citing works
On attention models for human activity recognition
Murahari, V. S., & Plötz, T. (2018) · Proceedings of the 2018 ACM International Symposium on Wearable Computers · doi:10.1145/3267242.3267287 · Cited by 157 · see citing works
Kwon, H., Tong, C., Haresamudram, H., Gao, Y., Abowd, G. D., Lane, N. D., & Plötz, T. (2020) · Proceedings of the ACM on Interactive, Mobile, Wearable and Ubiquitous Technologies · doi:10.1145/3411841 · Cited by 152 · see citing works
Flagged first, then by citations
Research works printed but not linked: 2
- [1][1] Antoine Bordes, Sumit Chopra, and Jason Weston. 2014. Question Answering With Subgraph Embeddings. In Proceedings of the 2014 Conference on Empirical Methods in Natural Language Processing. 615–620. http://doi.org/10.3115/v1/D14- 1067 · DOI
- [2][2] Andreas Bulling, Ulf Blanke, and Bernt Schiele. 2014. A Tutorial on Human Activity Recognition Using Body-Worn Inertial Sensors. Comput. Surveys 46, 3 (2014). https://doi.org/10.1145/2499621 · DOI · Paper page
Cited by
Sorted by citations
Ciliberto, M., Fortes Rey, V., Calatroni, A., Lukowicz, P., & Roggen, D. (2021) · Frontiers in Computer Science · doi:10.3389/fcomp.2021.792065 · Cited by 35 · see citing works
“…ontology-based activity recognition Noor et al. (2018), unsupervised domain adaptation Chang et al. (2020), zero-shot learning Wu et al. (2020), convolutional feature optimisation Hiremath and Ploetz (2021) and deep network architecture optimisation Bock et al. (2021); Pellatt and Roggen (2021).”
BackgroundHaresamudram, 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 · see citing works
“Bock et al. [16], however, found that using a shallower LSTM in the DeepConvLSTM setup is better for many HAR datasets.”
If only we had more data!: Sensor-Based Human Activity Recognition in Challenging Scenarios
Plötz, T. (2023) · 2023 IEEE International Conference on Pervasive Computing and Communications Workshops and other Affiliated Events (PerCom Workshops) · doi:10.1109/percomworkshops56833.2023.10150267 · Cited by 16 · see citing works
Haresamudram, H., Tang, C. I., Suh, S., Lukowicz, P., & Ploetz, T. (2024) · arXiv · Cited by 31 · see citing works
“Bock et al. [16], however, found that using a shallower LSTM in the DeepConvLSTM setup is better for many HAR datasets.”
- Authors
- Bock, M., Hölzemann, A., Moeller, M., & Van Laerhoven, K.
- Affiliations
- University of Siegen (DE)
- University of Siegen, Germany
- Venue
- 2021 International Symposium on Wearable Computers
- Publisher
- ACM
- Published
- 2021-09-21
- Pages
- 7-12
- OpenAlex
- W3193145553
- Cited by
- 74
- References
- 11
- Citations by year
- 2026: 5 · 2025: 19 · 2024: 22 · 2023: 17 · 2022: 9 · 2021: 2
- FWCI
- 4.09
- Type
- conference paper
- Language
- en
- Open access
- green
- arxiv.org
- Locations
- International Symposium on Wearable Computers (publishedVersion)
- arXiv (Cornell University) (submittedVersion)
- arXiv (Cornell University)
- 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 › Computer Science › Human-Computer Interaction › Hand Gesture Recognition Systems
- Keywords
- human activity recognition, sensor-based activity recognition, hidden units, training time reduction, sequential data
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.
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
