TSTELM: Two-Stage Transfer Extreme Learning Machine for Unsupervised Domain Adaptation.

As a single-layer feedforward network (SLFN), extreme learning machine (ELM) has been successfully applied for classification and regression in machine learning due to its faster training speed and better generalization. However, it will perform poorly for domain adaptation in which the distributions between training data and testing data are inconsistent. In this article, we propose a novel ELM called two-stage transfer extreme learning machine (TSTELM) to solve this problem. At the statistical matching stage, we adopt maximum mean discrepancy (MMD) to narrow the distribution difference of the output layer between domains. In addition, at the subspace alignment stage, we align the source and target model parameters, design target cross-domain mean approximation, and add the output weight approximation to further promote the knowledge transferring across domains. Moreover, the prediction of test sample is jointly determined by the ELM parameters generated at the two stages. Finally, we investigate the proposed approach in classification task and conduct experiments on four public domain adaptation datasets. The result indicates that TSTELM could effectively enhance the knowledge transfer ability of ELM with higher accuracy than other existing transfer and non-transfer classifiers.

A Novel Septoplasty Technique for Patients With Nasal Fractures.

ABSTRACT: Nasal structures have both functional and cosmetic significance. These structures maintain the shape of the nose and regulate the nasal airflow. During trauma, fractures of the nasal bone are frequently associated with nasal septum deviations. This can lead to the nasal bone collapsing and nasal obstruction. The septoplasty technique is a major surgical intervention to improve nasal obstructions, with the submucosal resection of the deviated septum. In the past, septoplasty was deferred until the nasal bone fracture was healed to reduce the postoperative risk of saddle-nose and flat nose deformities. Advances in technology have enabled surgeons to attempt septoplasty together with a closed reduction of the nasal bone fraction. It is most important to preserve the septal support structure during surgery. Hence, we advocate that the nasal septum be reset in the midline rather than removed, by modified endoscopic septoplasty.

Transfer Extreme Learning Machine with Output Weight Alignment.

Extreme Learning Machine (ELM) as a fast and efficient neural network model in pattern recognition and machine learning will decline when the labeled training sample is insufficient. Transfer learning helps the target task to learn a reliable model by using plentiful labeled samples from the different but relevant domain. In this paper, we propose a supervised Extreme Learning Machine with knowledge transferability, called Transfer Extreme Learning Machine with Output Weight Alignment (TELM-OWA). Firstly, it reduces the distribution difference between domains by aligning the output weight matrix of the ELM trained by the labeled samples from the source and target domains. Secondly, the approximation between the interdomain ELM output weight matrix is added to the objective function to further realize the cross-domain transfer of knowledge. Thirdly, we consider the objective function as the least square problem and transform it into a standard ELM model to be efficiently solved. Finally, the effectiveness of the proposed algorithm is verified by classification experiments on 16 sets of image datasets and 6 sets of text datasets, and the result demonstrates the competitive performance of our method with respect to other ELM models and transfer learning approach.

Correction to: Two different therapies for the middle turbinate during endoscopic sinus surgery for chronic rhinosinusitis.

In the original publication of the article, Fig. 1a was missing and caption of Fig. 1a was published as caption of Fig. 1b. The correct Fig. 1 and captions are provided below.

Two different therapies for the middle turbinate during endoscopic sinus surgery for chronic rhinosinusitis.

BACKGROUND: The population of patients with chronic rhinosinusitis (CRS) has greatly increased. When medical treatment fails, endoscopic sinus surgery (ESS) is considered. In the present study, the value of two different therapies for the middle turbinate to optimize surgical outcomes was observed. Our objective was to determine a better management approach for the middle turbinate to effectively preserve the middle turbinate mucosa and function and avoid complications following ESS, such as nasal adhesions on the operative side. METHODS: Thirty patients [group A; treated prior to 2015 (primary surgery)] undergoing resection of the middle turbinate bone during complete ESS for CRS and 30 patients [group B; treated after 2015 (later surgery)] undergoing middle turbinate preservation and multiapproach therapy during complete ESS for CRS were observed. Nasal cavities were compared using perioperative sinus endoscopy (POSE) and Lund-Kennedy (LKES) scores preoperatively and at 15 days, 2 months and 1 year after ESS. RESULTS: Preoperatively, the POSE (8.83 ± 3.81 vs 9.15 ± 3.85, p = 0.45, for groups A and B, respectively) and LKES (4.23 ± 0.74 vs 4.13 ± 0.70, p = 0.34) scores were similar between groups. In group A, anterior adhesions were reported on six sides of the middle turbinate, severe adhesions were observed on two sides, mild adhesions were observed on one side, and adhesions occurred on two sides during follow-up. After retreatment, adhesions were still observed on two sides at 1 year. In group B, only mild anterior adhesions were observed on two sides. There was no difference between group A and group B at 15 days, and the POSE (4.31 ± 1.19 vs 4.07 ± 1.42, p = 0.11, for groups A and B, respectively) and LKES (3.35 ± 0.82 vs 3.33 ± 0.90, p = 0.91) scores were similar between groups. There was no significant difference in LKES (0.22 ± 0.49 vs 0.10 ± 0.35, p = 0.15) scores at 1 year between the two groups. There was a significant difference in the nasal cavities between group A and group B at 2 months and 1 year, where group B showed a better endoscopic appearance than group A at 2 months and 1 year (with POSE scores of 3.48 ± 0.83 vs 2.43 ± 1.38 (p = 0.00) and LKES scores of 1.35 ± 0.86 vs 1.15 ± 0.90 (p = 0.02) at 2 months for groups A and B, respectively, and POSE scores of 1.00 ± 0.96 vs 0.62 ± 0.87 (p = 0.001) at 1 year for groups A and B, respectively). CONCLUSIONS: Our results show that middle turbinate preservation and combined therapy was a better ESS method for CRS. Multiapproach middle conchoplasty, which is predominately a submucoperiosteal surgery, can preserve more of the mucosa and functions of the middle turbinate. Unlike the single-approach middle conchoplasty described in previous research, multiapproach middle conchoplasty is achieved by combining a three-step surgical procedure ("surgery, packing and removal") with "cocktail-style" postoperative packing and removal.

A relationship between the V4+/V5+ ratio and the surface dispersion, surface acidity, and redox performance of V2O5-WO3/TiO2 SCR catalysts.

A series of V2O5-WO3/TiO2 catalysts with different vanadium loading amounts were prepared by an impregnation method and were characterized by XRD, Raman spectroscopy, XPS, DRIFTS, Py-DRIFTS, NH3-TPD, H2-TPR, etc. The results show that the catalytic activity is related to the ratio of V4+/V5+. The variation of the V4+/V5+ ratio caused by the different dispersion states of vanadia oxide leads to changes in the surface acidity and redox properties of the catalysts. As the V4+/V5+ ratio reaches the maximum value, the apparent activation energy (E a) required to form the transition state on the Brønsted acid sites is the lowest. Artificial regulation of vanadium loading to properly increase V4+/V5+ content may affect the interactions between V, W, O and Ti atoms, which enhances NH3-SCR reaction performance.