Domain adaptation via incremental confidence samples into classification

Abstract

To accurately recognize similar objects in different domains, the key for domain adaptation is to learn new metrics so as to minimize the discrepancy of two domains. Recent works utilize joint probability domain adaptation to tackle this problem but get poor performance due to poor discriminability or transferability of data sets. The inaccurate pseudo‐labeling in the feature subspace can lead to a chain reaction of errors during iterations, and varieties of the joint probability distribution values further aggravate the miscalculation. To cope with the above problems, this paper proposes a unified framework by introducing Incremental Confidence Samples into Classification (ICSC). ICSC includes both incrementally labeling and adaptively adjusting. With the increase of confidence samples in each iteration, incrementally labeling is used to reduce error accumulations and progressively guarantee good classification performance effectively. Moreover, the tradeoff weight between within‐class and between‐class distance is adaptively adjusted according to the importance of transferability and discriminability. Consequently, the discrepancy minimization of within‐class and the discrepancy maximization of between‐class are achieved. Extensive experiments on several benchmark data sets demonstrate the effectiveness of the proposed method over the state‐of‐the‐art methods.

Adaptive graph-guided joint soft clustering and distribution alignment for cross-load and cross-device rotating machinery fault transfer diagnosis

“…(10) PDALC [50]: PDALC integrates class discriminant information and joint distribution alignment to learn a shared subspace, while the label correction mechanism facilitates the alignment of conditional distribution. (11) ICSC [51]: ICSC learns a projection matrix to minimize probability distribution differences, reduce intra-class distances, and increase inter-class distances. Finally, incremental confidence samples are employed to reduce classification errors in each iteration.…”

Dual selections based knowledge transfer learning for cross-subject motor imagery EEG classification

“…This issue significantly affects the feature adaptation process. Previous studies (Wang and Breckon, 2020 ; Teng et al, 2022 ) have shown that gradually predicting pseudo-labels from high-quality to low-quality samples yields better results than directly predicting pseudo labels for all samples from the target domain during the feature adaptation process. Therefore, to ensure a well-performing iterative process of feature adaptation, we employ a selective pseudo-labeling strategy.…”

Rolling bearing fault diagnosis based on the fusion of sparse filtering and discriminative domain adaptation method under multi-channel data-driven

“…Once the balance parameters are not selected properly, errors can accumulate during the iteration process, ultimately affecting the model's ability to identify faults. Therefore, to avoid the negative impact of the fixed equilibrium parameter u on the fault diagnosis model, we use an adaptive adjustment method using the covariance matrix as a class distance measure [39] in this paper to solve such problems.…”