Energy vs. Noise: Towards Robust Temporal Action Localization in Open-World

Authors: Chenyu Mu, Jiahua Li, Kun Wei, Cheng Deng

AAAI 2025 | Venue PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental Experiments on THUMOS14 and Activity Net1.3 datasets show that EDMP effectively enhances the robustness of TAL models. ... Table 1: The main results on the THUMOS14 and Activity Net1.3 datasets. ... Ablation Study In this section, we conduct ablation studies on the THUMOS14 dataset and the Action Former baseline under the mixed noise σB σC = 0.5 50% condition.
Researcher Affiliation Academia School of Electronic Engineering, Xidian University, Xi an 710071, China EMAIL
Pseudocode No The paper describes methods using equations and textual explanations, but it does not contain any explicitly labeled pseudocode or algorithm blocks with structured steps.
Open Source Code Yes Code https://github.com/XD-mu/EDMP
Open Datasets Yes Our experiments are conducted on two datasets: THUMOS14 (Idrees et al. 2017) and Activity Net1.3 (Caba Heilbron et al. 2015).
Dataset Splits Yes THUMOS14 provides 413 untrimmed sports videos for 20 action categories, including 200 videos for training and 213 videos for testing... Activity Net 1.3 provides 10,024 training, 4,926 validation, and 5,044 test videos with 200 action classes.
Hardware Specification No The paper does not explicitly mention any specific hardware used for running its experiments (e.g., GPU models, CPU types, or cloud instance specifications).
Software Dependencies No The metamodel employs the Adam W optimizer with a learning rate of 10 5. While an optimizer is mentioned, there are no specific version numbers for software components like programming languages (e.g., Python), libraries (e.g., PyTorch, TensorFlow), or other dependencies.
Experiment Setup Yes The metamodel employs the Adam W optimizer with a learning rate of 10 5. The ESO has a batch length L = 3, and the hyperparameters κ, λenergy, and γ are set to 1, 1, and 0.5, respectively, from grid search results. Action Former and Temporal Maxer are trained for 30 and 10 epochs on THUMOS14 and Activity Net1.3, respectively, with a 5-epoch warmup. Tri Det is trained for 20 and 5 epochs on THUMOS14 and Activity Net1.3, with warmup periods of 20 and 10 epochs.