# Don't freeze: Finetune encoders for better Self-Supervised HAR

arXiv: https://arxiv.org/abs/2307.01168 · HTML: https://mlsanity.com/paper/2307.01168/

## Key facts

- Cited by 2 papers in the map (data through 14 Oct 2026).
- Builds on 10 of its 17 references that are in the map.
- First posted on arXiv 3 Jul 2023.

## Builds on (most cited first, 10 of 10)

- [Multi-task Self-Supervised Learning for Human Activity Detection](https://mlsanity.com/paper/1907.11879/) (2019)
- [Exploring Contrastive Learning in Human Activity Recognition for Healthcare](https://mlsanity.com/paper/2011.11542/) (2020)
- [Self-supervised Learning for Human Activity Recognition Using 700,000 Person-days of Wearable Data](https://mlsanity.com/paper/2206.02909/) (2022)
- [Assessing the State of Self-Supervised Human Activity Recognition using Wearables](https://mlsanity.com/paper/2202.12938/) (2022)
- [Contrastive Predictive Coding for Human Activity Recognition](https://mlsanity.com/paper/2012.05333/) (2020)
- [Protecting Sensory Data against Sensitive Inferences](https://mlsanity.com/paper/1802.07802/) (2018)
- [What Makes Good Contrastive Learning on Small-Scale Wearable-based Tasks?](https://mlsanity.com/paper/2202.05998/) (2022)
- [SelfHAR: Improving Human Activity Recognition through Self-training with Unlabeled Data](https://mlsanity.com/paper/2102.06073/) (2021)
- [IMUTube: Automatic Extraction of Virtual on-body Accelerometry from Video for Human Activity Recognition](https://mlsanity.com/paper/2006.05675/) (2020)
- [Investigating Enhancements to Contrastive Predictive Coding for Human Activity Recognition](https://mlsanity.com/paper/2211.06173/) (2022)

## Built on by (most cited first, 2 of 2)

- [PIM: Physics-Informed Multi-task Pre-training for Improving Inertial Sensor-Based Human Activity Recognition](https://mlsanity.com/paper/2503.17978/) (2025)
- [ALS-HAR: Harnessing Wearable Ambient Light Sensors to Enhance IMU-based Human Activity Recogntion](https://mlsanity.com/paper/2408.09527/) (2024)

## Abstract (arXiv)

Recently self-supervised learning has been proposed in the field of human activity recognition as a solution to the labelled data availability problem. The idea being that by using pretext tasks such as reconstruction or contrastive predictive coding, useful representations can be learned that then can be used for classification. Those approaches follow the pretrain, freeze and fine-tune procedure. In this paper we will show how a simple change - not freezing the representation - leads to substantial performance gains across pretext tasks. The improvement was found in all four investigated datasets and across all four pretext tasks and is inversely proportional to amount of labelled data. Moreover the effect is present whether the pretext task is carried on the Capture24 dataset or directly in unlabelled data of the target dataset.