Data Mining - Concepts and Techniques.

We will review some of the major results in random graphs and some of the more challenging open problems. We will cover algorithmic and structural questions. We will touch on newer models, including those related to the WWW.

Random graphs

“Unpaired Image-to-Image Translation using Cycle-Consistent Adversarial Networks”の学習報告

Social Network Analysis

Emotion and Adaptation

With the development of artificial intelligence and pattern recognition, facial expression recognition plays a more and more important role in intelligent human-computer interaction. In this paper, we present a model named K-order emotional intensity model (K-EIM) which is based on K-Means clustering. Different from other related works, the proposed approach can quantify emotional intensity in an unsupervised way. And then the output from K-EIM is encoded. The coding results are used for the dynamic facial expression recognition. The experiment is conducted on Cohn-Kanade facial expression database and the support vector machine classifier is used for facial expression classification. This method achieved a dynamic facial expression recognition accuracy of 88.32% which suggest that the proposed method shows better performance and proves its validity. Moreover, effect of different segments of emotional intensity is also discussed in the paper.

Dynamic facial expression recognition based on K-order emotional intensity model

Multi-reservoirs EEG signal feature sensing and recognition method based on generative adversarial networks

Knowledge graph (KG) entity typing aims at inferring possible missing entity type instances in KG, which is a very significant but still under-explored subtask of knowledge graph completion. In this paper, we propose a novel approach for KG entity typing which is trained by jointly utilizing local typing knowledge from existing entity type assertions and global triple knowledge in KGs. Specifically, we present two distinct knowledge-driven effective mechanisms of entity type inference. Accordingly, we build two novel embedding models to realize the mechanisms. Afterward, a joint model via connecting them is used to infer missing entity type instances, which favors inferences that agree with both entity type instances and triple knowledge in KGs. Experimental results on two real-world datasets (Freebase and YAGO) demonstrate the effectiveness of our proposed mechanisms and models for improving KG entity typing.

Knowledge graph entity typing via learning connecting embeddings

BeSense: Leveraging WiFi Channel Data and Computational Intelligence for Behavior

In order to recognize human emotion, a three-dimensional (3D) emotion space model was established and used to identify the text emotional orientation visually and intuitively. An emotion dictionary contains informations of eight basic emotions, was built based on the emotion corpus. By using this emotion dictionary, emotion keywords which behave distributions of 3D structure can be projected into the emotion space. Emotion distributions were transformed into an emotion matrix. By analyzing the emotion matrix, not only binary classification of texts but also multi-emotion attributes can be investigated. The best precision 91% of a binary classification is obtained in this work.

Fuji Ren

Papers

Dynamic facial expression recognition based on K-order emotional intensity model

Multi-reservoirs EEG signal feature sensing and recognition method based on generative adversarial networks

Knowledge graph entity typing via learning connecting embeddings

BeSense: Leveraging WiFi Channel Data and Computational Intelligence for Behavior

The study on text emotional orientation based on a three-dimensional emotion space model