A Comparative Study on the Integration of Attention Mechanisms in GAN Architectures
Bibliographic record
Abstract
To enhance the structural reconstruction capabilities and semantic consistency of generative adversarial networks (GANs) in high-resolution image generation, this study focuses on the integration methods and performance differences of various attention mechanisms within GAN architectures. A systematic analysis was conducted on four mainstream mechanisms—self-attention, SE, CBAM, and non-local—across the generator, discriminator, and bidirectional embedding paths. Using the COCO and CelebA-HQ datasets, with a unified image resolution of 256×256, controlled experiments were designed with parameter increases kept within ±10%. Evaluation metrics included inception score, FID, PSNR, SSIM, and loss variance. The results show that self-attention and non-local modules have significant advantages in modeling long-range dependencies and global semantics, with FID reduced to 41.5 and 39.8, PSNR improved to 26.9 dB and 27.1 dB, SSIM reaching 0.834 and 0.839, and training stability metrics such as loss variance reduced to 0.049 and 0.047. In contrast, SE and CBAM achieve performance improvements with extremely low parameter growth, making them suitable for model lightweight requirements. The dual-end embedding path performed optimally across all metrics, demonstrating the effectiveness of collaborative modeling between the generator and discriminator. Analysis suggests that different attention mechanisms significantly impact model performance, with integration methods and embedding positions determining the ability to restore image details and model semantic consistency. This provides theoretical support and experimental evidence for future optimization of attention mechanism structures and the development of dynamic integration strategies.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.002 | 0.004 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.001 | 0.000 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.001 | 0.001 |
| Open science | 0.001 | 0.001 |
| Research integrity | 0.001 | 0.001 |
| Insufficient payload (model declined to judge) | 0.001 | 0.000 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".