No abstract

No abstract

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.

No abstract

The widespread use of 3D models has become an interest in steganography. In this paper, we present a novel data hiding method for 3D models. Based on the representation information, the key idea is to consider the vertex index as a message block. A message consists of three types of message blocks, namely, unique, repeated and 1-bit or 0-bit repetitions blocks. Three embedding methods, namely, vertex index embedding (VIE), dynamic-length bit-string mapping (DBM), and repeated bits embedding (RBE), each best for a respective type of message block are devised. All message vertices are then arranged in the light of proposed vertex order and output as the stego-model. The message block is extracted from the vertex of the stego-model by the order of vertex sequence and its index in the ordered vertex list. While integrating the proposed methods our scheme gains high capacity compared to existing techniques while preserving reasonable robustness. In addition, our method is efficient, 25 times faster than previous techniques. With high capacity, several novel applications like content annotation for large documents, 3D meta model for related resources, etc., becomes possible, making the 3D model as an excellent data container.