Research Work (Master)

• Problems

Few Training and Dataset Shift (Training/Test) in Machine Learning and Data Mining

• Topics

Take use of knowledge extracted from test examples (Semi-supervised Learning)

Take use of knowledge extracted from training examples from other domains (Cross-domain Learning)

• Related work

Feature selection and extraction

Ensemble learning

Manifold learning

Distance metric learning

• Selected Publications (as First Author)

[1] Cross-domain Representation-learning Framework with Combination of Class-separate and Domain-merge Objectives. 18th SIGKDD Workshop, 2012 (Conf. Level: Rank 1, Workshop Level: ?)

[2] Dynamical ensemble learning with model-friendly classifiers for domain adaptation. 21st ICPR, 2012. (Conf. Level: Rank 2)

[3] Subject transfer framework. Neurocomputing. 2012. (Jour. Level: Leading)

[4] Transferable discriminative dimensionality reduction. 23rd ICTAI, 2011. (Conf. Level: Rank 2)

[5] Semi-supervised feature extraction with local temporal regularization. 21st ICNN/IJCNN, 2011. (Conf. Level: Rank 2)

[6] Semi-supervised feature extraction for EEG classification. Pattern Analysis and Applications, 2012. (Jour. Level: Leading)

[7] Importance weighted extreme energy ratio. 17th ICONIP, 2011. (Conf. Level: Rank 3)

[8] Spatial filter selection with Lasso. 6th ADMA, 2010. (Conf. Level: Rank 3)

# Ranking references:

Computer Science Conference Rankings and Computer Science Journal Rankings

# [0] ― under review

• Master Thesis

Domain Adaptation and User-Transfer based Personalization Applications

(In Chinese: 域適應算法以及基於用戶群體遷移的個性化機器學習應用)