Research map / Image generation / Autoregressive and latent generation
When Attributes Disagree: Gradient Conflict in Image Aesthetic Assessment
Ye Wang, Maocai Dai, Jiang Xie, Xiuli Bi, Fei Tao, Xiao Li, Hong Yu
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 error-aware reweighting (ablation) | EAMBNet · LAPIS High-M hard samples | MAE: .126 -> .091 (-27.8%) | Improved |
| add error-aware reweighting (ablation) | EAMBNet · LAPIS Serr hard samples | MAE: .143 -> .086 (-39.9%) | Improved |
| add error-aware reweighting (ablation) | EAMBNet · LAPIS Low-M samples | MAE: .082 -> .077 (-6.1%) | Improved |
| add all four AGREE modules (ablation) | EAMB-Net · AVA | SRCC: .702 -> .732 | Improved |
| add all four AGREE modules (ablation) | EAMB-Net · AVA | MAE: .420 -> .387 | Improved |
| add AGREE to EAT | EAT · AVA | SRCC: .752 -> .789 | Improved |
| add AGREE to EAT | EAT · LAPIS | SRCC: .802 -> .853 | Improved |
| add AGREE to HKD-IAA | HKD-IAA · TAD66K | SRCC: .395 -> .432 | Improved |
| add AGREE to EAT | EAT · AADB | SRCC: .601 -> .669 | Mixed |
| add attribute decoupling to AGREE | EAT · LAPIS gradient conflict | conflict ratio: reduces conflict ratio by 24% at epoch 42 | Improved |
| add decoupled adapters instead of equal-size shared capacity (ablation) | AGREE decoupling-only · parameter-matched control | performance | Improved |
| increase attribute set size K to five (ablation) | AGREE · attribute-granularity | performance | Improved |
Closest papers by meaning
- Fine-grained Image Aesthetic Assessment: Learning Discriminative Scores from Relative Ranks (CVPR 2026 · similarity 0.89)
- ArtiMuse: Fine-Grained Image Aesthetics Assessment with Joint Scoring and Expert-Level Understanding (CVPR 2026 · similarity 0.88)
- Advancing Comprehensive Aesthetic Insight with Multi-Scale Text-Guided Self-Supervised Learning (AAAI 2025 · similarity 0.87)
- ELTA: An Enhancer against Long-Tail for Aesthetics-oriented Models (ICML 2024 · similarity 0.86)
- Bridging Cognitive Gap: Hierarchical Description Learning for Artistic Image Aesthetics Assessment (AAAI 2026 · similarity 0.86)
- Unlocking the Essence of Beauty: Advanced Aesthetic Reasoning with Relative-Absolute Policy Optimization (ICLR 2026 · similarity 0.86)
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.
