# Matching Skeleton-based Activity Representations with Heterogeneous Signals for HAR

arXiv: https://arxiv.org/abs/2503.14547 · HTML: https://mlsanity.com/paper/2503.14547/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 18 of its 66 references that are in the map.
- First posted on arXiv 17 Mar 2025.

## Builds on (most cited first, 18 of 18)

- [Attention Is All You Need](https://mlsanity.com/paper/1706.03762/) (2017)
- [Deep Residual Learning for Image Recognition](https://mlsanity.com/paper/1512.03385/) (2015)
- [Llama 2: Open Foundation and Fine-Tuned Chat Models](https://mlsanity.com/paper/2307.09288/) (2023)
- [node2vec: Scalable Feature Learning for Networks](https://mlsanity.com/paper/1607.00653/) (2016)
- [Ensemble Distillation for Robust Model Fusion in Federated Learning](https://mlsanity.com/paper/2006.07242/) (2020)
- [T2M-GPT: Generating Human Motion from Textual Descriptions with Discrete Representations](https://mlsanity.com/paper/2301.06052/) (2023)
- [A Transformer-based Framework for Multivariate Time Series Representation Learning](https://mlsanity.com/paper/2010.02803/) (2020)
- [ROCKET: Exceptionally fast and accurate time series classification using random convolutional kernels](https://mlsanity.com/paper/1910.13051/) (2019)
- [What Makes Good Contrastive Learning on Small-Scale Wearable-based Tasks?](https://mlsanity.com/paper/2202.05998/) (2022)
- [Unsupervised Feature Learning of Human Actions as Trajectories in Pose Embedding Manifold](https://mlsanity.com/paper/1812.02592/) (2018)
- [IMUTube: Automatic Extraction of Virtual on-body Accelerometry from Video for Human Activity Recognition](https://mlsanity.com/paper/2006.05675/) (2020)
- [IMUGPT 2.0: Language-Based Cross Modality Transfer for Sensor-Based Human Activity Recognition](https://mlsanity.com/paper/2402.01049/) (2024)
- [Sense and Learn: Self-Supervision for Omnipresent Sensors](https://mlsanity.com/paper/2009.13233/) (2020)
- [Language-based Action Concept Spaces Improve Video Self-Supervised Learning](https://mlsanity.com/paper/2307.10922/) (2023)
- [Unleashing the Power of Shared Label Structures for Human Activity Recognition](https://mlsanity.com/paper/2301.03462/) (2023)
- [Self-Supervised Multimodal Fusion Transformer for Passive Activity Recognition](https://mlsanity.com/paper/2209.03765/) (2022)
- [Navigating Alignment for Non-identical Client Class Sets: A Label Name-Anchored Federated Learning Framework](https://mlsanity.com/paper/2301.00489/) (2023)
- [ReL-SAR: Representation Learning for Skeleton Action Recognition with Convolutional Transformers and BYOL](https://mlsanity.com/paper/2409.05749/) (2024)

## Abstract (arXiv)

In human activity recognition (HAR), activity labels have typically been encoded in one-hot format, which has a recent shift towards using textual representations to provide contextual knowledge. Here, we argue that HAR should be anchored to physical motion data, as motion forms the basis of activity and applies effectively across sensing systems, whereas text is inherently limited. We propose SKELAR, a novel HAR framework that pretrains activity representations from skeleton data and matches them with heterogeneous HAR signals. Our method addresses two major challenges: (1) capturing core motion knowledge without context-specific details. We achieve this through a self-supervised coarse angle reconstruction task that recovers joint rotation angles, invariant to both users and deployments; (2) adapting the representations to downstream tasks with varying modalities and focuses. To address this, we introduce a self-attention matching module that dynamically prioritizes relevant body parts in a data-driven manner. Given the lack of corresponding labels in existing skeleton data, we establish MASD, a new HAR dataset with IMU, WiFi, and skeleton, collected from 20 subjects performing 27 activities. This is the first broadly applicable HAR dataset with time-synchronized data across three modalities. Experiments show that SKELAR achieves the state-of-the-art performance in both full-shot and few-shot settings. We also demonstrate that SKELAR can effectively leverage synthetic skeleton data to extend its use in scenarios without skeleton collections.