# ASMa: Asymmetric Spatio-temporal Masking for Skeleton Action Representation Learning

arXiv: https://arxiv.org/abs/2602.06251 · HTML: https://mlsanity.com/paper/2602.06251/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 33 of its 47 references that are in the map.
- First posted on arXiv 5 Feb 2026.

## Builds on (most cited first, 30 of 33)

- [Attention Is All You Need](https://mlsanity.com/paper/1706.03762/) (2017)
- [A Simple Framework for Contrastive Learning of Visual Representations](https://mlsanity.com/paper/2002.05709/) (2020)
- [DINOv2: Learning Robust Visual Features without Supervision](https://mlsanity.com/paper/2304.07193/) (2023)
- [Momentum Contrast for Unsupervised Visual Representation Learning](https://mlsanity.com/paper/1911.05722/) (2019)
- [Masked Autoencoders Are Scalable Vision Learners](https://mlsanity.com/paper/2111.06377/) (2021)
- [Bootstrap your own latent: A new approach to self-supervised Learning](https://mlsanity.com/paper/2006.07733/) (2020)
- [Exploring Simple Siamese Representation Learning](https://mlsanity.com/paper/2011.10566/) (2020)
- [Dynamic Graph CNN for Learning on Point Clouds](https://mlsanity.com/paper/1801.07829/) (2018)
- [Barlow Twins: Self-Supervised Learning via Redundancy Reduction](https://mlsanity.com/paper/2103.03230/) (2021)
- [Unsupervised Representation Learning by Predicting Image Rotations](https://mlsanity.com/paper/1803.07728/) (2018)
- [Context Encoders: Feature Learning by Inpainting](https://mlsanity.com/paper/1604.07379/) (2016)
- [Spatial Temporal Graph Convolutional Networks for Skeleton-Based Action Recognition](https://mlsanity.com/paper/1801.07455/) (2018)
- [NTU RGB+D: A Large Scale Dataset for 3D Human Activity Analysis](https://mlsanity.com/paper/1604.02808/) (2016)
- [NTU RGB+D 120: A Large-Scale Benchmark for 3D Human Activity Understanding](https://mlsanity.com/paper/1905.04757/) (2019)
- [Disentangling and Unifying Graph Convolutions for Skeleton-Based Action Recognition](https://mlsanity.com/paper/2003.14111/) (2020)
- [Channel-wise Topology Refinement Graph Convolution for Skeleton-Based Action Recognition](https://mlsanity.com/paper/2107.12213/) (2021)
- [Revisiting Skeleton-based Action Recognition](https://mlsanity.com/paper/2104.13586/) (2021)
- [A New Representation of Skeleton Sequences for 3D Action Recognition](https://mlsanity.com/paper/1703.03492/) (2017)
- [Representation Learning by Learning to Count](https://mlsanity.com/paper/1708.06734/) (2017)
- [3D Human Action Representation Learning via Cross-View Consistency Pursuit](https://mlsanity.com/paper/2104.14466/) (2021)
- [MS2L: Multi-Task Self-Supervised Learning for Skeleton Based Action Recognition](https://mlsanity.com/paper/2010.05599/) (2020)
- [Contrastive Learning from Extremely Augmented Skeleton Sequences for Self-supervised Action Recognition](https://mlsanity.com/paper/2112.03590/) (2021)
- [Skeleton-Contrastive 3D Action Representation Learning](https://mlsanity.com/paper/2108.03656/) (2021)
- [Skeleton Cloud Colorization for Unsupervised 3D Action Representation Learning](https://mlsanity.com/paper/2108.01959/) (2021)
- [A Survey on 3D Skeleton-Based Action Recognition Using Learning Method](https://mlsanity.com/paper/2002.05907/) (2020)
- [SkeletonMAE: Graph-based Masked Autoencoder for Skeleton Sequence Pre-training](https://mlsanity.com/paper/2307.08476/) (2023)
- [CMD: Self-supervised 3D Action Representation Learning with Cross-modal Mutual Distillation](https://mlsanity.com/paper/2208.12448/) (2022)
- [SkeletonMAE: Spatial-Temporal Masked Autoencoders for Self-supervised Skeleton Action Recognition](https://mlsanity.com/paper/2209.02399/) (2022)
- [Self-supervised Action Representation Learning from Partial Spatio-Temporal Skeleton Sequences](https://mlsanity.com/paper/2302.09018/) (2023)
- [SCD-Net: Spatiotemporal Clues Disentanglement Network for Self-supervised Skeleton-based Action Recognition](https://mlsanity.com/paper/2309.05834/) (2023)

## Abstract (arXiv)

Self-supervised learning (SSL) has shown remarkable success in skeleton-based action recognition by leveraging data augmentations to learn meaningful representations. However, existing SSL methods rely on data augmentations that predominantly focus on masking high-motion frames and high-degree joints such as joints with degree 3 or 4. This results in biased and incomplete feature representations that struggle to generalize across varied motion patterns. To address this, we propose Asymmetric Spatio-temporal Masking (ASMa) for Skeleton Action Representation Learning, a novel combination of masking to learn a full spectrum of spatio-temporal dynamics inherent in human actions. ASMa employs two complementary masking strategies: one that selectively masks high-degree joints and low-motion, and another that masks low-degree joints and high-motion frames. These masking strategies ensure a more balanced and comprehensive skeleton representation learning. Furthermore, we introduce a learnable feature alignment module to effectively align the representations learned from both masked views. To facilitate deployment in resource-constrained settings and on low-resource devices, we compress the learned and aligned representation into a lightweight model using knowledge distillation. Extensive experiments on NTU RGB+D 60, NTU RGB+D 120, and PKU-MMD datasets demonstrate that our approach outperforms existing SSL methods with an average improvement of 2.7-4.4% in fine-tuning and up to 5.9% in transfer learning to noisy datasets and achieves competitive performance compared to fully supervised baselines. Our distilled model achieves 91.4% parameter reduction and 3x faster inference on edge devices while maintaining competitive accuracy, enabling practical deployment in resource-constrained scenarios.