# Multimodal Skeleton-Based Action Representation Learning via Decomposition and Composition

arXiv: https://arxiv.org/abs/2512.21064 · HTML: https://mlsanity.com/paper/2512.21064/

## Key facts

- Cited by 1 papers in the map (data through 14 Oct 2026).
- Builds on 39 of its 54 references that are in the map.
- First posted on arXiv 24 Dec 2025.

## Builds on (most cited first, 30 of 39)

- [Adam: A Method for Stochastic Optimization](https://mlsanity.com/paper/1412.6980/) (2014)
- [Masked Autoencoders Are Scalable Vision Learners](https://mlsanity.com/paper/2111.06377/) (2021)
- [Spatial Temporal Graph Convolutional Networks for Skeleton-Based Action Recognition](https://mlsanity.com/paper/1801.07455/) (2018)
- [VICReg: Variance-Invariance-Covariance Regularization for Self-Supervised Learning](https://mlsanity.com/paper/2105.04906/) (2021)
- [NTU RGB+D: A Large Scale Dataset for 3D Human Activity Analysis](https://mlsanity.com/paper/1604.02808/) (2016)
- [X3D: Expanding Architectures for Efficient Video Recognition](https://mlsanity.com/paper/2004.04730/) (2020)
- [NTU RGB+D 120: A Large-Scale Benchmark for 3D Human Activity Understanding](https://mlsanity.com/paper/1905.04757/) (2019)
- [Video Action Transformer Network](https://mlsanity.com/paper/1812.02707/) (2018)
- [MotionBERT: A Unified Perspective on Learning Human Motion Representations](https://mlsanity.com/paper/2210.06551/) (2022)
- [A New Representation of Skeleton Sequences for 3D Action Recognition](https://mlsanity.com/paper/1703.03492/) (2017)
- [3D Human Action Representation Learning via Cross-View Consistency Pursuit](https://mlsanity.com/paper/2104.14466/) (2021)
- [Skeleton-Based Action Recognition with Multi-Stream Adaptive Graph Convolutional Networks](https://mlsanity.com/paper/1912.06971/) (2019)
- [Skeleton-based Action Recognition via Spatial and Temporal Transformer Networks](https://mlsanity.com/paper/2008.07404/) (2020)
- [MS2L: Multi-Task Self-Supervised Learning for Skeleton Based Action Recognition](https://mlsanity.com/paper/2010.05599/) (2020)
- [PREDICT & CLUSTER: Unsupervised Skeleton Based Action Recognition](https://mlsanity.com/paper/1911.12409/) (2019)
- [Contrastive Learning from Extremely Augmented Skeleton Sequences for Self-supervised Action Recognition](https://mlsanity.com/paper/2112.03590/) (2021)
- [Constructing Stronger and Faster Baselines for Skeleton-based Action Recognition](https://mlsanity.com/paper/2106.15125/) (2021)
- [Skeleton-Contrastive 3D Action Representation Learning](https://mlsanity.com/paper/2108.03656/) (2021)
- [Learning Discriminative Representations for Skeleton Based Action Recognition](https://mlsanity.com/paper/2303.03729/) (2023)
- [Modeling Temporal Dynamics and Spatial Configurations of Actions Using Two-Stream Recurrent Neural Networks](https://mlsanity.com/paper/1704.02581/) (2017)
- [HYperbolic Self-Paced Learning for Self-Supervised Skeleton-based Action Representations](https://mlsanity.com/paper/2303.06242/) (2023)
- [Actionlet-Dependent Contrastive Learning for Unsupervised Skeleton-Based Action Recognition](https://mlsanity.com/paper/2303.10904/) (2023)
- [SkeletonMAE: Graph-based Masked Autoencoder for Skeleton Sequence Pre-training](https://mlsanity.com/paper/2307.08476/) (2023)
- [Masked Motion Predictors are Strong 3D Action Representation Learners](https://mlsanity.com/paper/2308.07092/) (2023)
- [3Mformer: Multi-order Multi-mode Transformer for Skeletal Action Recognition](https://mlsanity.com/paper/2303.14474/) (2023)
- [CMD: Self-supervised 3D Action Representation Learning with Cross-modal Mutual Distillation](https://mlsanity.com/paper/2208.12448/) (2022)
- [LLMs are Good Action Recognizers](https://mlsanity.com/paper/2404.00532/) (2024)
- [Contrastive Positive Mining for Unsupervised 3D Action Representation Learning](https://mlsanity.com/paper/2208.03497/) (2022)
- [Global-local Motion Transformer for Unsupervised Skeleton-based Action Learning](https://mlsanity.com/paper/2207.06101/) (2022)
- [Hierarchical Consistent Contrastive Learning for Skeleton-Based Action Recognition with Growing Augmentations](https://mlsanity.com/paper/2211.13466/) (2022)

## Built on by (most cited first, 1 of 1)

- [MoAKE: Toward Unified All-in-One Action Quality Assessment via Mixture of Action Knowledge Experts](https://mlsanity.com/paper/2607.19826/) (2026)

## Abstract (arXiv)

Multimodal human action understanding is a significant problem in computer vision, with the central challenge being the effective utilization of the complementarity among diverse modalities while maintaining model efficiency. However, most existing methods rely on simple late fusion to enhance performance, which results in substantial computational overhead. Although early fusion with a shared backbone for all modalities is efficient, it struggles to achieve excellent performance. To address the dilemma of balancing efficiency and effectiveness, we introduce a self-supervised multimodal skeleton-based action representation learning framework, named Decomposition and Composition. The Decomposition strategy meticulously decomposes the fused multimodal features into distinct unimodal features, subsequently aligning them with their respective ground truth unimodal counterparts. On the other hand, the Composition strategy integrates multiple unimodal features, leveraging them as self-supervised guidance to enhance the learning of multimodal representations. Extensive experiments on the NTU RGB+D 60, NTU RGB+D 120, and PKU-MMD II datasets demonstrate that the proposed method strikes an excellent balance between computational cost and model performance.