# Dynamic User-controllable Privacy-preserving Few-shot Sensing Framework

arXiv: https://arxiv.org/abs/2508.03989 · HTML: https://mlsanity.com/paper/2508.03989/

## Key facts

- Cited by 8 papers in the map (data through 14 Oct 2026).
- Builds on 16 of its 43 references that are in the map.
- First posted on arXiv 6 Aug 2025.

## Builds on (most cited first, 16 of 16)

- [Learning Transferable Visual Models From Natural Language Supervision](https://mlsanity.com/paper/2103.00020/) (2021)
- [Supervised Contrastive Learning](https://mlsanity.com/paper/2004.11362/) (2020)
- [Mobile Sensor Data Anonymization](https://mlsanity.com/paper/1810.11546/) (2018)
- [Contrastive Learning for Neural Topic Model](https://mlsanity.com/paper/2110.12764/) (2021)
- [Protecting Sensory Data against Sensitive Inferences](https://mlsanity.com/paper/1802.07802/) (2018)
- [IMUGPT 2.0: Language-Based Cross Modality Transfer for Sensor-Based Human Activity Recognition](https://mlsanity.com/paper/2402.01049/) (2024)
- [ZSTAD: Zero-Shot Temporal Activity Detection](https://mlsanity.com/paper/2003.05583/) (2020)
- [Federated Intrusion Detection for IoT with Heterogeneous Cohort Privacy](https://mlsanity.com/paper/2101.09878/) (2021)
- [Large Language Models are Zero-Shot Recognizers for Activities of Daily Living](https://mlsanity.com/paper/2407.01238/) (2024)
- [Few-Shot Learning-Based Human Activity Recognition](https://mlsanity.com/paper/1903.10416/) (2019)
- [A Comprehensive Review of Few-shot Action Recognition](https://mlsanity.com/paper/2407.14744/) (2024)
- [Cross-Domain HAR: Few Shot Transfer Learning for Human Activity Recognition](https://mlsanity.com/paper/2310.14390/) (2023)
- [PCAP-Backdoor: Backdoor Poisoning Generator for Network Traffic in CPS/IoT Environments](https://mlsanity.com/paper/2501.15563/) (2025)
- [Replacement AutoEncoder: A Privacy-Preserving Algorithm for Sensory Data Analysis](https://mlsanity.com/paper/1710.06564/) (2017)
- [Leveraging Foundation Models for Zero-Shot IoT Sensing](https://mlsanity.com/paper/2407.19893/) (2024)
- [You can't always get what you want: towards user-controlled privacy on Android](https://mlsanity.com/paper/2106.02483/) (2021)

## Built on by (most cited first, 8 of 8)

- [Contrastive Continual Learning for Model Adaptability in Internet of Things](https://mlsanity.com/paper/2602.04881/) (2026)
- [Backdoor Attacks on Multi-modal Contrastive Learning](https://mlsanity.com/paper/2601.11006/) (2026)
- [Contrastive Learning for Privacy Enhancements in Industrial Internet of Things](https://mlsanity.com/paper/2602.00515/) (2026)
- [Backdoor Attacks on Contrastive Continual Learning for IoT Systems](https://mlsanity.com/paper/2602.13062/) (2026)
- [CFD-HAR: User-controllable Privacy through Conditional Feature Disentanglement](https://mlsanity.com/paper/2603.11526/) (2026)
- [Backdoor Attacks on Fault Detection and Localization in Cyber-Physical Systems](https://mlsanity.com/paper/2605.27674/) (2026)
- [Continual Backdoor Training in IoT/CPS](https://mlsanity.com/paper/2606.14987/) (2026)
- [CLIP-guided Diffusion Model for Backdoor Generation in Sensor-based Human Activity Recognition](https://mlsanity.com/paper/2606.22837/) (2026)

## Abstract (arXiv)

User-controllable privacy is important in modern sensing systems, as privacy preferences can vary significantly from person to person and may evolve over time. This is especially relevant in devices equipped with Inertial Measurement Unit (IMU) sensors, such as smartphones and wearables, which continuously collect rich time-series data that can inadvertently expose sensitive user behaviors. While prior work has proposed privacy-preserving methods for sensor data, most rely on static, predefined privacy labels or require large quantities of private training data, limiting their adaptability and user agency. In this work, we introduce PrivCLIP, a dynamic, user-controllable, few-shot privacy-preserving sensing framework. PrivCLIP allows users to specify and modify their privacy preferences by categorizing activities as sensitive (black-listed), non-sensitive (white-listed), or neutral (gray-listed). Leveraging a multimodal contrastive learning approach, PrivCLIP aligns IMU sensor data with natural language activity descriptions in a shared embedding space, enabling few-shot detection of sensitive activities. When a privacy-sensitive activity is identified, the system uses a language-guided activity sanitizer and a motion generation module (IMU-GPT) to transform the original data into a privacy-compliant version that semantically resembles a non-sensitive activity. We evaluate PrivCLIP on multiple human activity recognition datasets and demonstrate that it significantly outperforms baseline methods in terms of both privacy protection and data utility.