Dr. Yu Wang


Statistical Laboratory
Department of Pure Math and Matematical Sciences
University of Cambridge
Cambridge, UK, CB3 0WB
Email: yw323 at cam.ac.uk


I am now a postdoctoral research associate in Stats-Lab, Department of Pure Math and Mathematical Statistics, University of Cambridge. I am co-supervised by Prof. John Aston and Dr. Carola-Bibiane Schönlieb. I achieved my PhD degree in Computer Science from University of Cambridge, UK in 2016, supervised by Dr. Ian Wassell.

My research mainly focuses on machine learning topics. Specifically, I actively research with variational Bayesian inference, deep learning, generative models, and sparsity learning.

During my PhD, I was extremingly keen in investigating the mysterious Bayesian algorithms by bridging them with the underlying closed form optimization formulations. I believe by understanding these hidden connections, boosting the model towards more successful inference and regression is possible. My research also pays much emphasis on sparsity and low rank driven machine learning phenomenon. Based on my research, I developed novel efficient statistical Bayesian models for clustering, recognition, classification and generative tasks.

One of my favorite work was to successfully derive the closed form sparsity driven cost function hidden behind the hierarchical statistical clustering techniques (published in ICML 2015). My latest (UAI 2017) work introduces deep recurrent network into the Variational auto encoder (VAE), which enables to recycle the dirty data (See the paper for what I mean)!

(I recently moved so I no longer update this website. )


Hidden Talents of the Variational Autoencoder

Bin Dai, Yu Wang, Gang Hua, John Aston, David Wipf

Journal of Machine Learning Research (JMLR), July, 2018

Connections with Robust PCA and the Role of Emergent Sparsity in Variational Autoencoder Models

Bin Dai, Yu Wang, Gang Hua, John Aston, David Wipf

ICML Workshop in Theoretical Foundations and Applications of Deep Generative Models (TADGM 2018), July, 2018

Green Generative Modeling: Recycling Dirty Data using Recurrent Variational Autoencoders

Yu Wang, Bin Dai, Gang Hua, John Aston, David Wipf

Uncertainty of Artifical Intellegnece (UAI), August, 2017

Sparse Bayesian Multitask Learning for Subspace Segmentation

Yu Wang, David Wipf, Jonh Aston

CVPR 2017 WiCV (Workshop short paper). July 2017

Structured sparsity learning -- Taming the sparsity under structure

Yu Wang

University of Cambridge Technical Report, August, 2016

Simultaneous Bayesian Sparse Approximation With Structured Sparse Models

Wei Chen, David Wipf, Yu Wang, Yang Liu, Ian Wassell

IEEE Transactions on Signal Processing (TSP), Dec 2016

Clustered Sparse Bayesian Learning

Yu Wang, David Wipf, Jeong Min Yun, Wei Chen, Ian J. Wassell

The Conference on Uncertainty in Artificial Intelligence (UAI), Jul 2015

Multi-Task Learning for Subspace Segmentation

Yu Wang, David Wipf, Qing Ling, Wei Chen, Ian Wassell

International Conference on Machine Learning (ICML), Jul 2015

Exploiting the Convex-Concave Penalty for Tracking: A Novel Dynamic Reweighted Sparse Bayesian Learning Algorithm

Yu Wang, David Wipf, Wei Chen, Ian Wassell

IEEE International Conference on Acoustics, Speech, and Signal Processing (ICASSP), May 2014

Exploiting Hidden Block Sparsity: Interdependent Matching Pursuit for Cyclic Feature Detection

Yu Wang, Wei Chen, Ian Wassell

IEEE Global Communications Conference (GLOBECOM), Dec 2013