| Autor: |
Rakesh Kumar, Ankith Jain |
| Jazyk: |
angličtina |
| Rok vydání: |
2018 |
| Předmět: |
|
| Zdroj: |
Rakesh Kumar, Ankith Jain. (2018). Statistical Analysis of WCET on DNN. UC Riverside: Electrical Engineering. Retrieved from: http://www.escholarship.org/uc/item/3vz734j8 |
| Popis: |
The current research work on determining the worst-case execution time (WCET)focuses mainly on real-time systems since this is a key parameter in evaluating the reliabilityof a time-critical entity. There is a real dearth of research in estimating WCET measurementsin the area of deep neural networks (DNN). This work proposes a novel approach thatpredicts the probabilistic WCET (pWCET) of DNN based image classication models suchas GoogleNet and CaeNet. The proposed approach uses actual measurement of the DNNstotal inference time that considers any variations in the input size and employs ExtremeValue Theory (EVT) to estimate the pWCET.The work also discusses a unique approach topredict the pWCET of image resizing given the variations in the input sizes of the imagesby estimating the pWCET of the single pixel and multiplying it with the actual image size.In addition to this, it achieves a condence level of 99% for its pWCET estimates. |
| Databáze: |
OpenAIRE |
| Externí odkaz: |
|
