# Dynamic Inference: A New Approach Toward Efficient Video Action Recognition

arXiv: https://arxiv.org/abs/2002.03342 · HTML: https://mlsanity.com/paper/2002.03342/

## Key facts

- Cited by 25 papers in the map (data through 7 Oct 2026).
- First posted on arXiv 9 Feb 2020.

## Built on by (most cited first, 25 of 25)

- [Dynamic Neural Networks: A Survey](https://mlsanity.com/paper/2102.04906/) (2021)
- [MoEfication: Transformer Feed-forward Layers are Mixtures of Experts](https://mlsanity.com/paper/2110.01786/) (2021)
- [MoViNets: Mobile Video Networks for Efficient Video Recognition](https://mlsanity.com/paper/2103.11511/) (2021)
- [Adaptive Focus for Efficient Video Recognition](https://mlsanity.com/paper/2105.03245/) (2021)
- [AdaFuse: Adaptive Temporal Fusion Network for Efficient Action Recognition](https://mlsanity.com/paper/2102.05775/) (2021)
- [Bidirectional Cross-Modal Knowledge Exploration for Video Recognition with Pre-trained Vision-Language Models](https://mlsanity.com/paper/2301.00182/) (2022)
- [ASCNet: Self-supervised Video Representation Learning with Appearance-Speed Consistency](https://mlsanity.com/paper/2106.02342/) (2021)
- [MAR: Masked Autoencoders for Efficient Action Recognition](https://mlsanity.com/paper/2207.11660/) (2022)
- [Temporal Action Proposal Generation with Transformers](https://mlsanity.com/paper/2105.12043/) (2021)
- [Temporal Action Proposal Generation with Background Constraint](https://mlsanity.com/paper/2112.07984/) (2021)
- [NSNet: Non-saliency Suppression Sampler for Efficient Video Recognition](https://mlsanity.com/paper/2207.10388/) (2022)
- [Computation-efficient Deep Learning for Computer Vision: A Survey](https://mlsanity.com/paper/2308.13998/) (2023)
- [DSANet: Dynamic Segment Aggregation Network for Video-Level Representation Learning](https://mlsanity.com/paper/2105.12085/) (2021)
- [What Can Simple Arithmetic Operations Do for Temporal Modeling?](https://mlsanity.com/paper/2307.08908/) (2023)
- [Temporal Saliency Query Network for Efficient Video Recognition](https://mlsanity.com/paper/2207.10379/) (2022)
- [Rethinking Resolution in the Context of Efficient Video Recognition](https://mlsanity.com/paper/2209.12797/) (2022)
- [Searching for Two-Stream Models in Multivariate Space for Video Recognition](https://mlsanity.com/paper/2108.12957/) (2021)
- [Good Practices and A Strong Baseline for Traffic Anomaly Detection](https://mlsanity.com/paper/2105.03827/) (2021)
- [Efficient Visual Recognition with Deep Neural Networks: A Survey on Recent Advances and New Directions](https://mlsanity.com/paper/2108.13055/) (2021)
- [Rapid Elastic Architecture Search under Specialized Classes and Resource Constraints](https://mlsanity.com/paper/2108.01224/) (2021)
- [GTM: Gray Temporal Model for Video Recognition](https://mlsanity.com/paper/2110.10348/) (2021)
- [Auto-X3D: Ultra-Efficient Video Understanding via Finer-Grained Neural Architecture Search](https://mlsanity.com/paper/2112.04710/) (2021)
- [Unifying Synergies between Self-supervised Learning and Dynamic Computation](https://mlsanity.com/paper/2301.09164/) (2023)
- [HaltingVT: Adaptive Token Halting Transformer for Efficient Video Recognition](https://mlsanity.com/paper/2401.04975/) (2024)
- [A Survey on Efficiency Optimization Techniques for DNN-based Video Analytics: Process Systems, Algorithms, and Applications](https://mlsanity.com/paper/2507.15628/) (2025)

## Abstract (arXiv)

Though action recognition in videos has achieved great success recently, it remains a challenging task due to the massive computational cost. Designing lightweight networks is a possible solution, but it may degrade the recognition performance. In this paper, we innovatively propose a general dynamic inference idea to improve inference efficiency by leveraging the variation in the distinguishability of different videos. The dynamic inference approach can be achieved from aspects of the network depth and the number of input video frames, or even in a joint input-wise and network depth-wise manner. In a nutshell, we treat input frames and network depth of the computational graph as a 2-dimensional grid, and several checkpoints are placed on this grid in advance with a prediction module. The inference is carried out progressively on the grid by following some predefined route, whenever the inference process comes across a checkpoint, an early prediction can be made depending on whether the early stop criteria meets. For the proof-of-concept purpose, we instantiate three dynamic inference frameworks using two well-known backbone CNNs. In these instances, we overcome the drawback of limited temporal coverage resulted from an early prediction by a novel frame permutation scheme, and alleviate the conflict between progressive computation and video temporal relation modeling by introducing an online temporal shift module. Extensive experiments are conducted to thoroughly analyze the effectiveness of our ideas and to inspire future research efforts. Results on various datasets also evident the superiority of our approach.