Non-parametric Bayesian Prior for Variational Auto-encoder
This is the final project for course Advanced Topics in Bayesian Statistics.
You can find the mid-term presentation about hierarchical dirichlet process here.
CNX's blog
This is the final project for course Advanced Topics in Bayesian Statistics.
You can find the mid-term presentation about hierarchical dirichlet process here.
This semester I took the Advanced Topics in Bayesian Statistics from Prof. Yanxun Xu. It is a interesting course and this is my presentation slides in the final term related to variational auto-encoder with non-parametric bayesian prior. Thie is the first part of this seris.
This semester I took the Advanced Topics in Bayesian Statistics from Prof. Yanxun Xu. It is a interesting course and this is my presentation slides in the mid term related to hierarchical dirichlet process:
Recent I have done some experiments using new models such as attention model for spoofing detection(still working on the BTAS challenge). It is hard to improve the result because the error rate for development set is quite small. But it should be useful for the speaker verification.
I firstly noticed this question when I worked as an intern in AISpeech.
At the end of 2015, let me start a brief summary for my 2015. This post started at 23:59, 31st Dec.
After reading this paper, I decided to implement the network and do some experiments on CIFAR-10 which is small enough. This is the first time for me to do vision tasks(besides MNIST which don’t require much knowledge on vision).
This reading list refered following articles:
CNCC2015 was held in Hefei from Oct. 22 to Oct.25. It was organized by China Computer Federation(CCF). I was invited to take part in CNCC2015 as one of the excellent college student.
Interspeech 2015 was held in ICD, Dresden, Germany. I took part in this conference because I had two papers published here. This is the first time to attend an international conference for me. The first part of this post includes lots of photos. LG Flex2 was used to take these photos. All photos are compressed using TinyPNG. ~78% space is saved. The second part includes analysis of some papers.