Adversarial Auxiliary Weighted Subdomain Adaptation for Open-Set Deep Transfer Bridge Damage Diagnosis.

Deep learning models have been widely used in data-driven bridge structural damage diagnosis methods in recent years. However, these methods require training and test datasets to satisfy the same distribution, which is difficult to satisfy in practice. Domain adaptation transfer learning is an efficient method to solve this problem. Most of the current domain adaptation methods focus on close-set scenarios with the same classes in the source and target domains. However, in practical applications, new damage caused by long-term degradation often makes the target and source domains dissimilar in the class space. For such challenging open-set scenarios, existing domain adaptation methods will be powerless. To effectively solve the above problems, an adversarial auxiliary weighted subdomain adaptation algorithm is proposed for open-set scenarios. Adversarial learning is introduced to proposed an adversarial auxiliary weighting scheme to reflect the similarity of target samples with source classes. It effectively distinguishes unknown damage from known states. This paper further proposes a multi-channel multi-kernel weighted local maximum mean discrepancy metric (MCMK-WLMMD) to capture the fine-grained transferable information for conditional distribution alignment (sub-domain alignment). Extensive experiments on transfer tasks between three bridges verify the effectiveness of the algorithm in open-set scenarios.

[A method to extract signals related to fatigue during visual display terminal (VDT) operation using independent component analysis].

With the recent, rapid developments of computers, adaptation between the human being and computer is being improved. On the other hand, mental fatigue is increased by continuous, long-time use. The problem of mental fatigue is now a new focus of research. This study proposes a method of evaluating mental fatigue. In this study, we used a method to extract signals related to fatigue using Independent Component Analysis (ICA). The subjects were 10 male students. After assuming a workload of simulated visual display terminal (VDT) operation for 30 min, the subjects sat in a relaxed position for 30 min. In this process, we measured work performance, subjective symptoms of feelings of fatigue and physiological measures. As a result, there was a high value of the correlation coefficient between these signals and the subjective symptoms of feeling of fatigue. We suggest that by using ICA we can determine the relationship between Independent Components and subjective feelings of fatigue for those under work stress, work strain and in a state of fatigue and stress. However, further experiments are needed to obtain conclusive findings of our research.

[Fatigue estimation of visual display terminal (VDT) workers by physiological data--preliminary study of a new computer program for VDT work analysis].

The methods for assessing VDT workload based on individual capabilities are necessary to evaluate general fatigue by VDT work. We have developed a new computer program for assessing data entry performance of each individual. Four healthy male volunteers conducted a data entry task at their own pace using this program under a controlled environment. They performed the task for 30 minutes twice with 10 minutes rest between them. As a result, the average number of inputs was 133.2 characters/min in the first task and 129.1 characters/min after taking the rest. The error rate was 8.6% and 10.6% in each task. This program was useful to evaluate their data entry performance. We also reported our concern with a drop in nasal skin temperature that was observed during these VDT tasks.