Computational Intelligence for Computer-Aided Design Machine Learning Techniques for Microcontrollers Performance Screenings
Year: 2024
Type: dissertation
Abstract: 3.1 Examples of activation functions commonly used in neural networks.37 3.2 Examples of loss functions utilized in neural network training.Here, y represents the predicted output of the network, while y ˆdenotes the true target values, and C is the number of classes. . . . . . . . .38 3.3 Example of confusion matrix . . . . . . . . . . . . . . . . . . . . . .44 5.1 Normalized root mean squared error (nRMSE) on the test set using basic models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .64 5.2 Basic models scores (mean values on 10 patterns) . . . . . . . . . .65 5.3 Comparison between PCA and UFS (mean nRMSE on 10 patterns, advanced independent models) . . . . . . . . . . . . . . . . . . . . .68 5.4 Advanced models vs Base models scores (mean values on 10 patterns, independent target) . . . . . . . . . . . . . . . . . . . . . . .69 5.5 Pearson correlation between Labels . . . . . . . . . . . . . . . . . .69 5.6 Comparison between advanced models in-chain or single target (P min target), average over 20 run. . . . . . . . . . . . . . . . . . . . . . .71 5.7 Area under the nRMSE curve for different algorithm (Polynomial Ridge, Random Forest, SVR) . . . . . . . . . . . . . . . . . . . . .72 5.8 Polynomial Ridge Regression Mean and variance nRMSE during active learning using all possible initial wafer choices . . . . . . . .74 5.9 Correlation between selection strategies and nRMSE on the f max Ridge regression models with polynomial features . . . . . . . . . .75 5.10 Prediction performances of P min target on test set with different methods for outlier detection on labels (Polynomial Ridge Regression) 78 6.1 Prediction performance metrics of deep feature extractors on product A1 . . . . . . . . . . . . .
Source: PORTO@Iris
Author: BELLARMINO, NICOLO'
Language: English
Subfield: Materials Chemistry
Field: Materials Science
Domain: Physical Sciences
Open Access status: green
Funder: Infineon Technologies
OpenAlex ID: W7217034482