Mukherjee, Subhaditya (2023) Proxy Attention : Approximating Attention in CNNs using Gradient Based Techniques. Master's Thesis / Essay, Artificial Intelligence.
|
Text
Subhaditya_final_thesis_report.pdf Download (60MB) | Preview |
|
![[img]](https://fse.studenttheses.ub.rug.nl/style/images/fileicons/text.png) |
Text
toestemming.pdf Restricted to Registered users only Download (128kB) |
Abstract
The problem statement of this study is creating a novel augmentation technique Proxy Attention, that uses attention maps to improve the performance of any model by de-weighting the regions that are not important for the classification task and guiding the model to better focus on critical parts of the image. In turn, this method should also improve the explainability of the model while maintaining the same architecture and hyperparam- eters
| Item Type: | Thesis (Master's Thesis / Essay) |
|---|---|
| Supervisor name: | Valdenegro Toro, M.A. and Mohades Kasaei, S.H. |
| Degree programme: | Artificial Intelligence |
| Thesis type: | Master's Thesis / Essay |
| Language: | English |
| Date Deposited: | 24 Jul 2023 11:29 |
| Last Modified: | 24 Jul 2023 11:29 |
| URI: | https://fse.studenttheses.ub.rug.nl/id/eprint/30889 |
Actions (login required)
 |
View Item |