# Learn to cycle: Time-consistent feature discovery for action recognition

arXiv: https://arxiv.org/abs/2006.08247 · HTML: https://mlsanity.com/paper/2006.08247/

## Key facts

- Cited by 8 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 15 Jun 2020.

## Built on by (most cited first, 8 of 8)

- [InternVideo: General Video Foundation Models via Generative and Discriminative Learning](https://mlsanity.com/paper/2212.03191/) (2022)
- [VATT: Transformers for Multimodal Self-Supervised Learning from Raw Video, Audio and Text](https://mlsanity.com/paper/2104.11178/) (2021)
- [MoViNets: Mobile Video Networks for Efficient Video Recognition](https://mlsanity.com/paper/2103.11511/) (2021)
- [AdaPool: Exponential Adaptive Pooling for Information-Retaining Downsampling](https://mlsanity.com/paper/2111.00772/) (2021)
- [Refining activation downsampling with SoftPool](https://mlsanity.com/paper/2101.00440/) (2021)
- [Can An Image Classifier Suffice For Action Recognition?](https://mlsanity.com/paper/2106.14104/) (2021)
- [Unsupervised Domain Adaptation for Video Transformers in Action Recognition](https://mlsanity.com/paper/2207.12842/) (2022)
- [Deep Neural Networks in Video Human Action Recognition: A Review](https://mlsanity.com/paper/2305.15692/) (2023)

## Abstract (arXiv)

Generalizing over temporal variations is a prerequisite for effective action recognition in videos. Despite significant advances in deep neural networks, it remains a challenge to focus on short-term discriminative motions in relation to the overall performance of an action. We address this challenge by allowing some flexibility in discovering relevant spatio-temporal features. We introduce Squeeze and Recursion Temporal Gates (SRTG), an approach that favors inputs with similar activations with potential temporal variations. We implement this idea with a novel CNN block that uses an LSTM to encapsulate feature dynamics, in conjunction with a temporal gate that is responsible for evaluating the consistency of the discovered dynamics and the modeled features. We show consistent improvement when using SRTG blocks, with only a minimal increase in the number of GFLOPs. On Kinetics-700, we perform on par with current state-of-the-art models, and outperform these on HACS, Moments in Time, UCF-101 and HMDB-51.