Tsung-Hsien Wen

I've been worked on natural language processing, speech processing for almost 6 years. I did my master in statistical dialogue systems and recurrent neural network language models. Now I'm studying my PhD in University of Cambridge, United Kingdom. My research focus on applying statistical methods and machine learning on dialogue systems. My current research project involves applied recurrent neural networks for stochastically natural language generation.

I did my master degree in National Taiwan University. My thesis is about using recurrent neural network language model to improve personalised speech recogniser. I'm now studying in University of Cambridge, UK, working on some projects related to language modeling, natural language generation and statistical dialogue systems.

I've worked with document classification in a Taiwan start-up company. I applied several statistical models to learn automatic classification of online webpages by looking at their contents. I applied both supervised methods (naive bayes, maximum entropy classifier, and SVM) and unsupervised methods (Latent Dirichlet Allocation, Probabilistic Latent Semantic Analysis). Now I'm studying in University of Cambridge, UK, on statistical dialogue systems and natural language generation.

I've been worked on applied machine learning for several years. I mainly worked on speech, text, and language related applications, with some experiences in image processing and computer vision. I'm now studying PhD in University of Cambridge, focus on statistical dialogue systems and natural language generation.