Fault injection attacks on SoftMax function in deep neural networks
Autor: | Shivam Bhasin, Dirmanto Jap, Yoo-Seung Won |
---|---|
Rok vydání: | 2021 |
Předmět: |
010302 applied physics
Artificial neural network Computer science business.industry Activation function Pattern recognition 02 engineering and technology Fault injection 01 natural sciences 020202 computer hardware & architecture 0103 physical sciences Softmax function 0202 electrical engineering electronic engineering information engineering Probability distribution Deep neural networks Fault analysis Artificial intelligence Layer (object-oriented design) business |
Zdroj: | CF |
Popis: | Softmax is commonly used activation function in neural networks to normalize the output to probability distribution over predicted classes. Being often deployed in the output layer, it can potentially be targeted by fault injection attacks to create misclassification. In this extended abstract, we perform a preliminary fault analysis of Softmax against single bit faults. |
Databáze: | OpenAIRE |
Externí odkaz: |