Search

Unsupervised domain adaptation (DA) aims to perform classification tasks on the target domain by leveraging rich labeled data in the existing source domain. The key insight of DA is to reduce domain divergence by learning domain-invariant features or transferable instances. Despite its rapid development, there still exist several challenges to explore. At the feature level, aligning both domains only in a single way (i.e., geometrical or statistical) has limited ability to reduce the domain divergence. At the instance level, interfering instances often obstruct learning a discriminant subspace when performing the geometrical alignment. At the classifier level, only minimizing the empirical risk on the source domain may result in a negative transfer. To tackle these challenges, this article proposes a novel DA method, called discriminant geometrical and statistical alignment (DGSA). DGSA first aligns the geometrical structure of both domains by projecting original space into a Grassmann manifold, then matches the statistical distributions of both domains by minimizing their maximum mean discrepancy on the manifold. In the former step, DGSA only selects the density peaks to learn the Grassmann manifold and so to reduce the influences of interfering instances. In addition, DGSA exploits the high-confidence soft labels of target landmarks to learn a more discriminant manifold. In the latter step, a structural risk minimization (SRM) classifier is learned to match the distributions (both marginal and conditional) and predict the target labels at the same time. Extensive experiments on objection recognition and human activity recognition tasks demonstrate that DGSA can achieve better performance than the comparison methods. [ABSTRACT FROM AUTHOR]

This article proposes a portable thermoelectric canteen system used for generating electricity while heating. A new step-shape bottom is designed, helping the canteen to be more evenly heated. The proposed canteen is a medium power and temperature difference (MPTD) system, whose main challenge is to keep system always operate at the maximum power point (MPP) when the temperature changes fast. To tackle this problem, this article proposed a new algorithm structure considered the transient response of the current. When an abnormal change in the current is obtained, $X$ is multiplied with the voltage at time $t$ , helping the operation point to relocate near the MPP quickly. The new algorithm structure can be applied with most existing thermoelectric maximum power point tracking (MPPT) algorithms, increasing the tracking efficiency. Furthermore, this article proposes an MPTD-specialized MPPT algorithm based on the new structure. The new algorithm uses the derivative of the power with respect to voltage and current to multiply the reference step $S$. When the temperature difference is 38  $^\circ$ C, the highest power of 0.7469 V and 70.0667 mW, respectively, can be achieved through the MPTD-specialized method. The tracking efficiency is 42.51% higher compared with the classic perturb & observe (P&O) algorithm. [ABSTRACT FROM AUTHOR]