# Knowledge-guided Disentanglement with Atomic Actions for Action Recognition

arXiv: https://arxiv.org/abs/2607.26097 · HTML: https://mlsanity.com/paper/2607.26097/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 36 of its 60 references that are in the map.
- First posted on arXiv 28 Jul 2026.

## Builds on (most cited first, 30 of 36)

- [Quo Vadis, Action Recognition? A New Model and the Kinetics Dataset](https://mlsanity.com/paper/1705.07750/) (2017)
- [Estimating or Propagating Gradients Through Stochastic Neurons for Conditional Computation](https://mlsanity.com/paper/1308.3432/) (2013)
- [ViViT: A Video Vision Transformer](https://mlsanity.com/paper/2103.15691/) (2021)
- [VideoMAE: Masked Autoencoders are Data-Efficient Learners for Self-Supervised Video Pre-Training](https://mlsanity.com/paper/2203.12602/) (2022)
- [Two-Stream Convolutional Networks for Action Recognition in Videos](https://mlsanity.com/paper/1406.2199/) (2014)
- [Vision Transformers Need Registers](https://mlsanity.com/paper/2309.16588/) (2023)
- [CLIP4Clip: An Empirical Study of CLIP for End to End Video Clip Retrieval](https://mlsanity.com/paper/2104.08860/) (2021)
- [Hollywood in Homes: Crowdsourcing Data Collection for Activity Understanding](https://mlsanity.com/paper/1604.01753/) (2016)
- [Rethinking Spatiotemporal Feature Learning: Speed-Accuracy Trade-offs in Video Classification](https://mlsanity.com/paper/1712.04851/) (2017)
- [Convolutional Two-Stream Network Fusion for Video Action Recognition](https://mlsanity.com/paper/1604.06573/) (2016)
- [Expanding Language-Image Pretrained Models for General Video Recognition](https://mlsanity.com/paper/2208.02816/) (2022)
- [ActionCLIP: A New Paradigm for Video Action Recognition](https://mlsanity.com/paper/2109.08472/) (2021)
- [Prompting Visual-Language Models for Efficient Video Understanding](https://mlsanity.com/paper/2112.04478/) (2021)
- [Action Genome: Actions as Composition of Spatio-temporal Scene Graphs](https://mlsanity.com/paper/1912.06992/) (2019)
- [Videos as Space-Time Region Graphs](https://mlsanity.com/paper/1806.01810/) (2018)
- [Frozen CLIP Models are Efficient Video Learners](https://mlsanity.com/paper/2208.03550/) (2022)
- [Fine-tuned CLIP Models are Efficient Video Learners](https://mlsanity.com/paper/2212.03640/) (2022)
- [Something-Else: Compositional Action Recognition with Spatial-Temporal Interaction Networks](https://mlsanity.com/paper/1912.09930/) (2019)
- [Spatial-Temporal Transformer for Dynamic Scene Graph Generation](https://mlsanity.com/paper/2107.12309/) (2021)
- [Actor-Context-Actor Relation Network for Spatio-Temporal Action Localization](https://mlsanity.com/paper/2006.07976/) (2020)
- [VindLU: A Recipe for Effective Video-and-Language Pretraining](https://mlsanity.com/paper/2212.05051/) (2022)
- [MultiSports: A Multi-Person Video Dataset of Spatio-Temporally Localized Sports Actions](https://mlsanity.com/paper/2105.07404/) (2021)
- [PSGAN: Pose and Expression Robust Spatial-Aware GAN for Customizable Makeup Transfer](https://mlsanity.com/paper/1909.06956/) (2019)
- [S3VAE: Self-Supervised Sequential VAE for Representation Disentanglement and Data Generation](https://mlsanity.com/paper/2005.11437/) (2020)
- [A Structured Model For Action Detection](https://mlsanity.com/paper/1812.03544/) (2018)
- [Spatiotemporal Pyramid Network for Video Action Recognition](https://mlsanity.com/paper/1903.01038/) (2019)
- [Explainable Video Action Reasoning via Prior Knowledge and State Transitions](https://mlsanity.com/paper/1908.10700/) (2019)
- [VicTR: Video-conditioned Text Representations for Activity Recognition](https://mlsanity.com/paper/2304.02560/) (2023)
- [Temporal Reasoning Graph for Activity Recognition](https://mlsanity.com/paper/1908.09995/) (2019)
- [OmniVid: A Generative Framework for Universal Video Understanding](https://mlsanity.com/paper/2403.17935/) (2024)

## Abstract (arXiv)

Action recognition in complex scenes often involves multiple concurrent fine-grained actions, making it challenging to model internal action structures. Most existing methods rely on holistic representations, which are insufficient for capturing subtle interactions and fine-grained semantics. While recent prompt-based approaches introduce disentanglement, they lack explicit semantic guidance, and methods based solely on visual or structured cues remain coarse-grained. In this paper, we propose Knowledge-guided Disentanglement with Atomic Actions (KDA), which leverages fine-grained semantic knowledge to enhance action representations and enable more precise disentanglement. Specifically, we use Large Language Models (LLMs) to decompose action labels into atomic actions, providing explicit spatial-temporal semantics. A Knowledge Injection Module (KIM) first integrates atomic action knowledge into video features. Based on this enhanced representation, a Knowledge Disentanglement Module (KDM) further disentangles atomic action knowledge to produce more precise semantic guidance for action disentanglement. A Knowledge Disentanglement Loss (KD Loss) is introduced to encourage clearer disentanglement of knowledge components within KDM. Extensive experiments demonstrate that KDA improves feature discriminability and achieves state-of-the-art performance on multi-label action recognition benchmarks. Moreover, KIM and KDM can be readily integrated into other methods, demonstrating strong generality.