Error Adaptive Classifier Boosting (EACB): Leveraging Data-Driven Training Towards Hardware Resilience for Signal Inference.
Zhuo WangRobert E. SchapireNaveen VermaPublished in: IEEE Trans. Circuits Syst. I Regul. Pap. (2015)
Keyphrases
- data driven
- training error
- training process
- boosted classifiers
- training set
- adaboost algorithm
- generalization error
- error rate
- discriminative classifiers
- training samples
- weak classifiers
- weak learners
- learning algorithm
- training examples
- ensemble learning
- low cost
- feature selection
- face detection
- training data
- object detection
- boosting algorithms
- structured prediction
- multiclass classification
- classifier training
- early stopping
- hearing aids
- fault tolerance
- decision stumps
- training phase
- multi layer perceptron
- text classifiers
- real time
- improving classification accuracy
- lower error rates
- accurate classifiers
- support vector
- error detection
- supervised learning
- linear classifiers
- signal processing
- base classifiers
- support vector machine
- bayesian networks
- minimum error
- decision trees
- hardware implementation
- partially labeled data
- neural network
- machine learning
- svm classifier
- class labels
- supervised training
- missing data
- multiple classifier systems
- ensemble classifier
- feature space
- active learning
- bayesian classifier