Deep neural networks (DNNs) are becoming prevalent, often outperforming manually-created systems. Unfortunately, DNN models are opaque to humans, and may behave in unexpected ways when deployed. One approach for allowing safer deployment of DNN models calls for augmenting them with hand-crafted override rules, which serve to override decisions made by the DNN model when certain criteria are met. Here, we propose to bring together DNNs and the well-studied scenario-based modeling paradigm, by expressing these override rules as simple and intuitive scenarios. This approach can lead to override rules that are comprehensible to humans, but are also sufficiently expressive and powerful to increase the overall safety of the model. We describe how to extend and apply scenario-based modeling to this new setting, and demonstrate our proposed technique on multiple DNN models.
|Original language||American English|
|Title of host publication||MODELSWARD 2020 - Proceedings of the 8th International Conference on Model-Driven Engineering and Software Development|
|Editors||Slimane Hammoudi, Luis Ferreira Pires, Bran Selic|
|Number of pages||11|
|State||Published - 2020|
|Event||8th International Conference on Model-Driven Engineering and Software Development, MODELSWARD 2020 - Valletta, Malta|
Duration: 25 Feb 2020 → 27 Feb 2020
|Name||MODELSWARD 2020 - Proceedings of the 8th International Conference on Model-Driven Engineering and Software Development|
|Conference||8th International Conference on Model-Driven Engineering and Software Development, MODELSWARD 2020|
|Period||25/02/20 → 27/02/20|
Bibliographical noteFunding Information:
We thank Yafim (Fima) Kazak for his contributions to this project, and the anonymous reviewers for their insightful comments. The project was partially supported by grants from the Binational Science Foundation (2017662) and the Israel Science Foundation (683/18).
Copyright © 2020 by SCITEPRESS - Science and Technology Publications, Lda. All rights reserved.
- Behavioral Programming
- Deep Neural Networks
- Machine Learning
- Scenario-based Modeling