Research map / Time series / Long-term forecasting
Convolution-Enhanced Evolving Attention Networks
Yujing Wang, Yaming Yang, Zhuo Li, Jiangang Bai, Mingliang Zhang, Xiangtai Li, Jing Yu, Ce Zhang, Gao Ming Huang, Yunhai Tong
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| add evolving attention (residual convolution over attention maps) | AA-ResNet-34 · ImageNet val | top-1 accuracy: 74.33 -> 75.23 | Improved |
| add evolving attention | AA-ResNet-152 · ImageNet val | top-1 accuracy: 79.10 -> 79.63 | Improved |
| add evolving attention | ViT/L-16 · ImageNet val | top-1 accuracy: 85.01 -> 85.17 | Improved |
| add evolving attention | BERT-Base · GLUE dev | average score: 83.1 -> 84.4 | Improved |
| add evolving attention | RoBERTa-Large · GLUE dev | average score: 88.5 -> 89.4 | Improved |
| add evolving attention | Transformer-Base · WMT'14 En-De | BLEU: 27.47 -> 27.56 | Improved |
| add evolving attention | Transformer-Base · WMT'14 En-Fr | BLEU: 40.79 -> 41.54 | Improved |
| add evolving attention and dilated convolution | Transformer · time-series classification | average accuracy: 0.679 -> 0.706 | Improved |
| add convolution module to attention evolution (ablation) | EA-AA-ResNet-34 · ImageNet val | top-1 accuracy: 74.34 -> 75.23 | Improved |
| add skip connection between attention maps (ablation) | EA-AA-ResNet-34 · ImageNet val | top-1 accuracy: 74.29 -> 75.23 | Improved |
| use 3x3 instead of 1x1 convolution (ablation) | EA-BERT-Base · GLUE CoLA | Matthews correlation: 57.5 -> 61.6 | Improved |
| add convolution module (ablation) | EA-BERT-Base · GLUE CoLA | Matthews correlation: 52.1 -> 61.6 | Improved |
Closest papers by meaning
- Rethinking Attention: Exploring Shallow Feed-Forward Neural Networks as an Alternative to Attention Layers in Transformers (Student Abstract) (AAAI 2024 · similarity 0.86)
- A-VL: Adaptive Attention for Large Vision-Language Models (AAAI 2025 · similarity 0.85)
- Unveiling and Harnessing Hidden Attention Sinks: Enhancing Large Language Models without Training through Attention Calibration (ICML 2024 · similarity 0.85)
- Transformers Learn Nonlinear Features In Context: Nonconvex Mean-field Dynamics on the Attention Landscape (ICML 2024 · similarity 0.85)
- Where Does Vision Meet Language? Understanding and Refining Visual Fusion in MLLMs via Contrastive Attention (CVPR 2026 · similarity 0.85)
- Large Vision-Language Models Get Lost in Attention (ICML 2026 · similarity 0.85)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
