摘要 :
Success of machine learning algorithms hinges on access to labeled dataset. Obtaining a labeled dataset is an expensive, challenging and time-consuming process, leading to the development of transfer learning (TL) methodology. TL ...
展开
Success of machine learning algorithms hinges on access to labeled dataset. Obtaining a labeled dataset is an expensive, challenging and time-consuming process, leading to the development of transfer learning (TL) methodology. TL incorporates gained knowledge from a previously trained source model into specific yet similar task models with limited data domain coverage. In this paper, we propose an automated targeted transfer learning (ATTL) method to resolve the transferability between source and target with minimal data requirements. The ATTL method decides how much target data is essential for model training, along with selected source data, to obtain the skateholder's specified performance metrics. The ATTL framework optimizes the system to select minimal target data based on two approaches: combinatorial coverage and adaptive selection methodology, along with specific source data for fine-tuning given a pre-trained source model. We evaluated the ATTL method on the Kaggle's ‘planes in satellite imagery’ dataset and the results identified that acquiring a small number of intentionally well-chosen samples from the target environment can achieve model performance of 97% in comparison to the baseline transfer learning accuracy of 92%.
收起
摘要 :
Success of machine learning algorithms hinges on access to labeled dataset. Obtaining a labeled dataset is an expensive, challenging and time-consuming process, leading to the development of transfer learning (TL) methodology. TL ...
展开
Success of machine learning algorithms hinges on access to labeled dataset. Obtaining a labeled dataset is an expensive, challenging and time-consuming process, leading to the development of transfer learning (TL) methodology. TL incorporates gained knowledge from a previously trained source model into specific yet similar task models with limited data domain coverage. In this paper, we propose an automated targeted transfer learning (ATTL) method to resolve the transferability between source and target with minimal data requirements. The ATTL method decides how much target data is essential for model training, along with selected source data, to obtain the skateholder's specified performance metrics. The ATTL framework optimizes the system to select minimal target data based on two approaches: combinatorial coverage and adaptive selection methodology, along with specific source data for fine-tuning given a pre-trained source model. We evaluated the ATTL method on the Kaggle's ‘planes in satellite imagery’ dataset and the results identified that acquiring a small number of intentionally well-chosen samples from the target environment can achieve model performance of 97% in comparison to the baseline transfer learning accuracy of 92%.
收起
摘要 :
In this paper, a physical unclonable function (PUF)-advanced encryption standard (AES)-PUF is proposed as a new PUF architecture by embedding an AES cryptographic circuit between two conventional PUF circuits to conceal their chal...
展开
In this paper, a physical unclonable function (PUF)-advanced encryption standard (AES)-PUF is proposed as a new PUF architecture by embedding an AES cryptographic circuit between two conventional PUF circuits to conceal their challenge-to-response pairs (CRPs) against machine learning attacks. Moreover, an internal confidential data is added to the secret key of the AES cryptographic circuit in the new PUF architecture to update the secret key in real-time against side-channel attacks. As shown in the results, even if 1 million number of data are enabled by the adversary to implement machine learning or side-channel attacks, the proposed PUF can not be cracked. By contrast, only 100,000 (1,000) number of data are sufficient to leak the confidential information of a conventional PUF via machine learning (side-channel) attacks.
收起
摘要 :
In this paper, a physical unclonable function (PUF)-advanced encryption standard (AES)-PUF is proposed as a new PUF architecture by embedding an AES cryptographic circuit between two conventional PUF circuits to conceal their chal...
展开
In this paper, a physical unclonable function (PUF)-advanced encryption standard (AES)-PUF is proposed as a new PUF architecture by embedding an AES cryptographic circuit between two conventional PUF circuits to conceal their challenge-to-response pairs (CRPs) against machine learning attacks. Moreover, an internal confidential data is added to the secret key of the AES cryptographic circuit in the new PUF architecture to update the secret key in real-time against side-channel attacks. As shown in the results, even if 1 million number of data are enabled by the adversary to implement machine learning or side-channel attacks, the proposed PUF can not be cracked. By contrast, only 100,000 (1,000) number of data are sufficient to leak the confidential information of a conventional PUF via machine learning (side-channel) attacks.
收起