# C-AAE: Compressively Anonymizing Autoencoders for Privacy-Preserving Activity Recognition in Healthcare Sensor Streams

arXiv: https://arxiv.org/abs/2507.18072 · HTML: https://mlsanity.com/paper/2507.18072/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 8 of its 31 references that are in the map.
- First posted on arXiv 24 Jul 2025.

## Builds on (most cited first, 8 of 8)

- [Unsupervised Representation Learning with Deep Convolutional Generative Adversarial Networks](https://mlsanity.com/paper/1511.06434/) (2015)
- [Semi-supervised Knowledge Transfer for Deep Learning from Private Training Data](https://mlsanity.com/paper/1610.05755/) (2016)
- [Mobile Sensor Data Anonymization](https://mlsanity.com/paper/1810.11546/) (2018)
- [Privacy-Preserving Adversarial Networks](https://mlsanity.com/paper/1712.07008/) (2017)
- [Privacy-Aware Eye Tracking Using Differential Privacy](https://mlsanity.com/paper/1812.08000/) (2018)
- [Privacy and Utility Preserving Sensor-Data Transformations](https://mlsanity.com/paper/1911.05996/) (2019)
- [Human Activity Recognition with Convolutional Neural Netowrks](https://mlsanity.com/paper/1906.01935/) (2019)
- [Privacy-Aware Human Mobility Prediction via Adversarial Networks](https://mlsanity.com/paper/2201.07519/) (2022)

## Abstract (arXiv)

Wearable accelerometers and gyroscopes encode fine-grained behavioural signatures that can be exploited to re-identify users, making privacy protection essential for healthcare applications. We introduce C-AAE, a compressive anonymizing autoencoder that marries an Anonymizing AutoEncoder (AAE) with Adaptive Differential Pulse-Code Modulation (ADPCM). The AAE first projects raw sensor windows into a latent space that retains activity-relevant features while suppressing identity cues. ADPCM then differentially encodes this latent stream, further masking residual identity information and shrinking the bitrate. Experiments on the MotionSense and PAMAP2 datasets show that C-AAE cuts user re-identification F1 scores by 10-15 percentage points relative to AAE alone, while keeping activity-recognition F1 within 5 percentage points of the unprotected baseline. ADPCM also reduces data volume by roughly 75 %, easing transmission and storage overheads. These results demonstrate that C-AAE offers a practical route to balancing privacy and utility in continuous, sensor-based activity recognition for healthcare.