Papers › Adaptive Parametric Activation

Adaptive Parametric Activation

11 Jul 2024arXiv:2407.08567archive 2025-07-28

Konstantinos Panagiotis Alexandridis, Jiankang Deng, Anh Nguyen, Shan Luo

The activation function plays a crucial role in model optimisation, yet the optimal choice remains unclear. For example, the Sigmoid activation is the de-facto activation in balanced classification tasks, however, in imbalanced classification, it proves inappropriate due to bias towards frequent classes. In this work, we delve deeper in this phenomenon by performing a comprehensive statistical analysis in the classification and intermediate layers of both balanced and imbalanced networks and we empirically show that aligning the activation function with the data distribution, enhances the performance in both balanced and imbalanced tasks. To this end, we propose the Adaptive Parametric Activation (APA) function, a novel and versatile activation function that unifies most common activation functions under a single formula. APA can be applied in both intermediate layers and attention layers, significantly outperforming the state-of-the-art on several imbalanced benchmarks such as ImageNet-LT, iNaturalist2018, Places-LT, CIFAR100-LT and LVIS and balanced benchmarks such as ImageNet1K, COCO and V3DET. The code is available at https://github.com/kostas1515/AGLU.

PaperPDFCode

In Syntology Open this paper in Syntology's Atlas, the map of the papers in Syntology's graph and their citations.

Code

kostas1515/aglu officialmentioned in papermentioned on GitHubpytorch report

Repository list and official/mentioned flags are the archive's, frozen 2025-07-28. Reachability, where shown, is from one Syntology probe window (2026-09-16 to 2026-09-18); repositories not probed show nothing. GitHub stars are not tracked.

Code Syntology ran Syntology

Not run by Syntology. Nothing on this page verifies that the listed code works.

Tasks

Instance SegmentationLong-tail Learningimbalanced classification

Results from the paper archive 2025-07-28

TaskDatasetModelMetricValueRank at snapshotLeaderboardReport
Instance Segmentation LVIS v1.0 val SE-R101-FPN-MaskRCNN-APA mask AP 30.7 #15 of 25 Archive leaderboard report
Instance Segmentation LVIS v1.0 val SE-R101-FPN-MaskRCNN-APA mask APr 23.6 #15 of 25 Archive leaderboard report
Instance Segmentation LVIS v1.0 val SE-R50-FPN-MaskRCNN-APA mask AP 29.1 #16 of 25 Archive leaderboard report
Instance Segmentation LVIS v1.0 val SE-R50-FPN-MaskRCNN-APA mask APr 21.6 #16 of 25 Archive leaderboard report
Long-tail Learning ImageNet-LT APA (SE-ResNext-50) Top-1 Accuracy 59.1 #17 of 69 Archive leaderboard report
Long-tail Learning ImageNet-LT APA (SE-ResNet-50) Top-1 Accuracy 57.9 #25 of 69 Archive leaderboard report
Long-tail Learning Places-LT APA (SE-ResNet-50) Top-1 Accuracy 42.0 #11 of 29 Archive leaderboard report
Long-tail Learning iNaturalist 2018 APA (SE-ResNet-50) Top-1 Accuracy 74.8 #16 of 43 Archive leaderboard report

Ranks are positions in the archive's leaderboards as they stood at the 2025-07-28 snapshot. Results published since then are not among these rows, so a rank here is not a current standing.

Methods

Introduced by this paper: AGLU

AGLUAPAAttentionSigmoid ActivationSoftmax

Report a problem or propose a change · a person checks every report against the paper or source before anything changes; decisions are listed on /corrections