Chi-Hsin Yu

Exploring Ensemble of Models in Taxonomy-based Cross-Domain Sentiment Classification

Most cross-domain sentiment classification techniques consider a domain as a whole set of opinionated instances for training. However, many online shopping websites organize their data in terms of taxonomy. With multiple domains (or, nodes) organized in a tree-structured representation, we propose a general ensemble algorithm which takes into account: 1) the model application, 2) the model weight and 3) the strategies for selecting the most related models with respect to a target node. The traditional sentiment classification technique SVM and the transfer learning algorithm Spectral Features Alignment (SFA) were applied as our model applications. In addition, the model weight takes the tree information and the similarity between domains into account. Finally, two strategies, cosine function and taxonomy-based regression model (TBRM) are proposed to select the most related models with respect to a target node. Experimental results showed both (cosine function and TBRM) proposed strategies outperform two baselines on an Amazon dataset. Three tasks of the proposed methods surpass the gold standard generated by the in-domain classifiers trained on the labeled data from the target nodes. Good results from the three tasks enable this algorithm to shed some new light on eliminating the major difficulties in transfer learning research: the distribution gap.

Using autocloning effects to develop broad-bandwidth, omnidirectional antireflection structures for silicon solar cells

In this study, we used the autocloning effect on pyramid structures to develop broad-bandwidth, omnidirectional antireflection structures for silicon solar cells. The angular dependence of reflectance on several pyramid structures was systematically investigated. The deposition of three-layer autocloned films reduced the refractive index gap between air and silicon, resulting in an increase in the amount of transmitted light and a decrease in the total light escaping. The average reflectance decreased dramatically to ca. 2-3% at incident angles from 0 to 60° for both sub-wavelength- and micrometer-scale pyramid structures. The measured reflectance of the autocloned structure was less than 4% in the wavelength range from 400 to 1000 nm for incident angles from 0 to 60°. Therefore, the autocloning technique, combined with optical thin films and optical gradient structures, is a practical and compatible method for the fabrication of broad-bandwidth, omnidirectional antireflection structures on silicon solar cells.

Taxonomy-based regression model for cross-domain sentiment classification

Most cross-domain sentiment classification techniques consider a domain as a whole set of instances for training. However, many online shopping websites organize their data in terms of taxonomy. This paper takes Amazon shopping website as an example, and proposes a tree-structured domain representation scheme in which each node in the tree is encoded as a bit sequence to preserve its relationship with all the other nodes in the tree. To select an appropriate source node for training in the domain taxonomy, we propose a Taxonomy-Based Regression Model (TBRM) which predicts the accuracy loss from multiple source nodes to a target node using the tree-structured domain representation combined with domain similarity and domain complexity. The source node with the smallest accuracy loss is used to train a classifier which makes a prediction on the target node. The results show that our TBRM achieves better performance than the regression models without considering the taxonomy information.

Using self-assembled nanoparticles to fabricate and optimize sub-wavelength textured structures in solar cells

In this study, we demonstrated the textured structure on silicon surface by metal assisted etching method, using Au nanoparticles as catalysts in HF and H2O2 solution. The size and density of the nanoparticles could be tuned easily. The porous layers filled with cylinder- or cone-shaped were uniformly formed by immersing the gold deposited silicon wafers in a mixed solution containing HF and H2O2 under different etching conditions. The optimized textured structure was close-packed pyramids-like surface in subwavelength scale and showed the lowest reflectance less than 0.5% over whole visible and near IR wavelengths. The large reduction of reflectance was attributed from the gradient refractive index of the silicon surface with the depth along the light propagation.