Counterfactual Attention for Facial Image Super-Resolution
Why this work is in the frame
A frame that forgets how it found something cannot be audited. These are the routes that admitted this work.
Bibliographic record
Abstract
Super-resolution (SR) is the task of recovering High-Resolution (HR) images from given Low- Resolution (LR) Images. Various SR methods are available in the literature. The attention mechanism is one of the widely used approaches in the field of SR. In this paper, Counterfactual Attention Learning (CAL) based on causal inference is applied to increase the quality of attention in Face Super-Resolution (FSR). This approach helps to assess the quality of attention and provides a strong signal to supervise the learning activity. The paper discusses the effect of the learned attention on the task of SR through counterfactual intervention and the effect is maximized to make the model learn useful attention for FSR. The effectiveness of the method is tested, and upscaling is achieved using the Scale-Arbitrary SR model (ArbSR), which can handle both integer and non-integer scale factors. The experiments are carried out for different scale factors on the CelebA dataset. The results show that the technique enhances the performance of FSR task by both quality of the image and PSNR.
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.
Full frame distilled prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. Learned from the 10,348 direct Codex labels and 10,348 direct Gemma labels. Candidate is the union of thresholded teacher heads; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels or direct frontier model labels.
Codex and Gemma teacher scores by category
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.000 | 0.000 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.000 | 0.000 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.000 | 0.001 |
| Open science | 0.000 | 0.000 |
| Research integrity | 0.000 | 0.000 |
| Insufficient payload (model declined to judge) | 0.000 | 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 it