Recently, various Deep Neural Network (DNN) models have been proposed for environments like embedded systems with stringent energy constraints. The fundamental problem of determining the ro-bustness of a DNN with respect to its energy consumption (energy robustness) is relatively unexplored compared to accuracy-based ro-bustness. This work investigates the energy robustness of Adaptive Neural Networks (AdNNs), a type of energy-saving DNNs proposed for many energy-sensitive domains and have recently gained traction. We propose EREBA, the first black-box testing method for determining the energy robustness of an AdNN. EREBA explores and infers the relationship between inputs and the energy con-sumption of AdNN s to generate energy surging samples. Extensive implementation and evaluation using three state-of-the-art AdNNs demonstrate that test inputs generated by EREBA could degrade the performance of the system substantially. The test inputs gener-ated by EREBA can increase the energy consumption of AdNN s by 2,000% compared to the original inputs. Our results also show that test inputs generated via EREBA are valuable in detecting energy surging inputs.

@inproceedings{HaqueY0L22,
  title = {EREBA},
  author = {Mirazul Haque and Yaswanth Yadlapalli and Wei Yang and Cong Liu},
  booktitle = {ICSE},
  year = {2022},
  doi = {10.1145/3510003.3510088}
}