Training multi-source domain adaptation network by mutual information estimation and minimization.

We address the problem of Multi-Source Domain Adaptation (MSDA), which trains a neural network using multiple labeled source datasets and an unlabeled target dataset, and expects the trained network to well classify the unlabeled target data. The main challenge in this problem is that the datasets are generated by relevant but different joint distributions. In this paper, we propose to address this challenge by estimating and minimizing the mutual information in the network latent feature space, which leads to the alignment of the source joint distributions and target joint distribution simultaneously. Here, the estimation of the mutual information is formulated into a convex optimization problem, such that the global optimal solution can be easily found. We conduct experiments on several public datasets, and show that our algorithm statistically outperforms its competitors. Video and code are available at https://github.com/sentaochen/Mutual-Information-Estimation-and-Minimization.

[Fabrication of alginate microsphere for controlled release and investigation of its release characteristics in vitro].

This study sought to producte alginate sodium microsphere for controlled release bovine serum albumin(BSA) and to investigate the protein release profile of the BSA-alginate sodium microsphere in vitro, which threw some light on the angiogenesis of tissue engineering bone with vascular endothelial growth factor (VEGF) controlled release under stress. The BSA-alginate sodium microsphere was fabricated with W/O emulsification and ion cross-linking method using alginate sodium. The appearance, microsphere diameter and envelopment rate were detected, and the release characteristics of the BSA-alginate sodium microsphere in vitro was investigated. The alginate microsphere was found to be spherical in shape and evenly distributed. Its mean grain diameter was determined to be 230 +/- 60 microm, carrying capacity 80.3 microg/mg and envelopment rate 61%. Smooth controlled release in BSA-alginate sodium microsphere was shown to last more than 2 weeks. Alginate sodium proved an excellent biodegradable material for protein or polypeptide controlled release. The emulsification and ion cross-linking method was noted to be simple; it was propitious to the structural and functional stablility of protein or polypeptide, thus leading to the prolonged efficacious time of the microsphere.