Summary of Explainability-driven Leaf Disease Classification Using Adversarial Training and Knowledge Distillation, by Sebastian-vasile Echim et al.
Explainability-Driven Leaf Disease Classification Using Adversarial Training and Knowledge Distillation
by Sebastian-Vasile Echim, Iulian-Marius Tăiatu, Dumitru-Clementin Cercel, Florin Pop
First submitted to arxiv on: 30 Dec 2023
Categories
- Main: Computer Vision and Pattern Recognition (cs.CV)
- Secondary: Machine Learning (cs.LG)
GrooveSquid.com Paper Summaries
GrooveSquid.com’s goal is to make artificial intelligence research accessible by summarizing AI papers in simpler terms. Each summary below covers the same AI paper, written at different levels of difficulty. The medium difficulty and low difficulty versions are original summaries written by GrooveSquid.com, while the high difficulty version is the paper’s original abstract. Feel free to learn from the version that suits you best!
| Summary difficulty | Written by | Summary |
|---|---|---|
| High | Paper authors | High Difficulty Summary Read the original abstract here |
| Medium | GrooveSquid.com (original content) | Medium Difficulty Summary The proposed research in plant leaf disease classification tackles three key areas: adversarial training, model explainability, and model compression. To enhance robustness against attacks, adversarial training is employed to ensure accurate classification even in the presence of threats. Explainability techniques are used to gain insights into the decision-making process, improving trust and transparency. Model compression methods optimize computational efficiency while maintaining performance. The results show that robustness can be traded for accuracy on a benchmark dataset, with performance reductions of 3%-20% for regular tests and gains of 50%-70% for adversarial attacks. A student model is also demonstrated to be 15-25 times more computationally efficient with a slight performance reduction, distilling the knowledge of more complex models. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary This research helps us understand how to better detect plant leaf diseases using special computer models. The goal is to make sure these models are good at recognizing real diseases and not fooled by fake ones. To do this, scientists use three important techniques: making the model stronger against attacks, explaining why it makes certain decisions, and making it more efficient with less power needed. By doing this, they found that the model can be a bit less accurate for normal tests but much better at detecting diseases when attacked. They also showed that a simpler model can do the same job as a more complex one, just using less energy. |
Keywords
* Artificial intelligence * Classification * Model compression * Student model
