Enhancing Explainability in Convolutional Neural Networks Using Entropy-Based Class Activation Maps
Bibliographic record
Abstract
With the emergence of visual sensors and their widespread application in intelligent systems, precise and interpretable visual explanations have become essential for ensuring the reliability and effectiveness of these systems. Sensor data, such as that from cameras operating in different spectra, LiDAR, or other imaging modalities, are often processed using complex deep learning methods, whose decision-making processes can be unclear. Accurate interpretation of network decisions is particularly critical in domains such as autonomous vehicles, medical imaging, and security systems. Moreover, during the development and deployment of deep learning architectures, the ability to accurately interpret results is crucial for identifying and mitigating any sources of bias in the training data, thereby ensuring fairness and robustness in the model’s performance. Explainable AI (XAI) techniques have garnered significant interest for their ability to reveal the rationale behind network decisions. In this work, we propose leveraging entropy information to enhance Class Activation Maps (CAMs). We explore two novel approaches: the first replaces the traditional gradient averaging scheme with entropy values to generate feature map weights, while the second directly utilizes entropy to weigh and sum feature maps, thereby reducing reliance on gradient-based methods, which can sometimes be unreliable. Our results demonstrate that entropy-based CAMs offer significant improvements in highlighting relevant regions of the input across various scenarios.
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 | 0.001 | 0.006 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.001 | 0.000 |
| Science and technology studies | 0.000 | 0.001 |
| Scholarly communication | 0.001 | 0.002 |
| Open science | 0.001 | 0.002 |
| Research integrity | 0.001 | 0.002 |
| 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".