# Efficient Deep Clustering of Human Activities and How to Improve Evaluation

arXiv: https://arxiv.org/abs/2209.08335 · HTML: https://mlsanity.com/paper/2209.08335/

## Key facts

- Cited by 5 papers in the map (data through 14 Oct 2026).
- Builds on 14 of its 47 references that are in the map.
- First posted on arXiv 17 Sep 2022.

## Builds on (most cited first, 14 of 14)

- [Adam: A Method for Stochastic Optimization](https://mlsanity.com/paper/1412.6980/) (2014)
- [UMAP: Uniform Manifold Approximation and Projection for Dimension Reduction](https://mlsanity.com/paper/1802.03426/) (2018)
- [Deep Learning for Sensor-based Activity Recognition: A Survey](https://mlsanity.com/paper/1707.03502/) (2017)
- [Deep Learning for Sensor-based Human Activity Recognition: Overview, Challenges and Opportunities](https://mlsanity.com/paper/2001.07416/) (2020)
- [Deep, Convolutional, and Recurrent Models for Human Activity Recognition using Wearables](https://mlsanity.com/paper/1604.08880/) (2016)
- [Protecting Sensory Data against Sensitive Inferences](https://mlsanity.com/paper/1802.07802/) (2018)
- [N2D: (Not Too) Deep Clustering via Clustering the Local Manifold of an Autoencoded Embedding](https://mlsanity.com/paper/1908.05968/) (2019)
- [Deep Activity Recognition Models with Triaxial Accelerometers](https://mlsanity.com/paper/1511.04664/) (2015)
- [Deep Clustering with a Dynamic Autoencoder: From Reconstruction towards Centroids Construction](https://mlsanity.com/paper/1901.07752/) (2019)
- [Deep Recurrent Neural Network for Mobile Human Activity Recognition with High Throughput](https://mlsanity.com/paper/1611.03607/) (2016)
- [Human Activity Recognition using Recurrent Neural Networks](https://mlsanity.com/paper/1804.07144/) (2018)
- [Adversarial Deep Feature Extraction Network for User Independent Human Activity Recognition](https://mlsanity.com/paper/2110.12163/) (2021)
- [Selective Pseudo-label Clustering](https://mlsanity.com/paper/2107.10692/) (2021)
- [An Unsupervised Approach for Automatic Activity Recognition based on Hidden Markov Model Regression](https://mlsanity.com/paper/1312.6965/) (2013)

## Built on by (most cited first, 5 of 5)

- [A Modular Approach for Multimodal Summarization of TV Shows](https://mlsanity.com/paper/2403.03823/) (2024)
- [Hard Regularization to Prevent Deep Online Clustering Collapse without Data Augmentation](https://mlsanity.com/paper/2303.16521/) (2023)
- [MAC-Gaze: Motion-Aware Continual Calibration for Mobile Gaze Tracking](https://mlsanity.com/paper/2505.22769/) (2025)
- [Towards a Universal Method for Meaningful Signal Detection](https://mlsanity.com/paper/2408.00016/) (2024)
- [Reducing Label Dependency in Human Activity Recognition with Wearables: From Supervised Learning to Novel Weakly Self-Supervised Approaches](https://mlsanity.com/paper/2512.19713/) (2025)

## Abstract (arXiv)

There has been much recent research on human activity re-cog-ni-tion (HAR), due to the proliferation of wearable sensors in watches and phones, and the advances of deep learning methods, which avoid the need to manually extract features from raw sensor signals. A significant disadvantage of deep learning applied to HAR is the need for manually labelled training data, which is especially difficult to obtain for HAR datasets. Progress is starting to be made in the unsupervised setting, in the form of deep HAR clustering models, which can assign labels to data without having been given any labels to train on, but there are problems with evaluating deep HAR clustering models, which makes assessing the field and devising new methods difficult. In this paper, we highlight several distinct problems with how deep HAR clustering models are evaluated, describing these problems in detail and conducting careful experiments to explicate the effect that they can have on results. We then discuss solutions to these problems, and suggest standard evaluation settings for future deep HAR clustering models. Additionally, we present a new deep clustering model for HAR. When tested under our proposed settings, our model performs better than (or on par with) existing models, while also being more efficient and better able to scale to more complex datasets by avoiding the need for an autoencoder.