Short Bio

I am a Ph.D. Candidate major in Electrical Engineering and minor in Computer Science at University of Southern California. I am lucky to have Prof. Gerard Medioni as my advisor and work on learning algorithms for structured data such as multivaraite time series. My research lab is Institute for Robotics and Intelligent Systems, Viterbi School of Engineering.

I worked as a research intern at SONY US Research Center for kernel metric learning and ranking, and a research intern at Microsoft Research Asia for graphical model based feature learning. I got my BS in Electronic Engineering from Tsinghua University in 2004.

My general passion is to design effecient and elegant solution to solve challenging problems in real world by a combination of machine learning, optimization and statistics.

News 

07/2011   One paper on Structured Time Series Alignment is accepted by ICCV 2011.

06/2011   Research intern at SONY US Research Center, San Jose, CA.

04/2011   Passed my PhD Oral Qualifying Examination.

06/2010   Attended CVPR 2010, San Franciso, CA.

05/2010   Attended AI & Statstics 2010 (first time AISTATS held in Europe), Sardinia, Italy. Thanks for the AISTATS travel award.

02/2010   One paper on nonparametric denoising for structured data is accepted by AISTATS 2010.

Research Interests 

1. On the theoretical side, my current research focuses on analytic analysis for Robust Manifold Learning and Tensor Voting. These are unsupervised learning methods for modeling high dimensional data and multivariate time series.

2. At the algorithm level, I am working on different machine learning domains include, modeling structured multivariate time series, spatio-temporal alignment, change-point detection, multi-manifold structure learning, metric and kernel learning, non-parametric manifold denoising, etc.

3. As an application, I am using the spatial-temporal manifold model for human motion analysis and recognition. I am also working on image contour grouping for image enhancement.

DMW is an unsupervised similarity learning and alignment algorithm for structured multivariate time series. Under the spatio-temporal manifold model, DMW can align two series with different length, dimensionality and sampling frequency.

 


 


Related Publications:

Dian Gong and Gerard Medioni, "Dynamic Manifold Warping for View Invariant Action Recognition", Proc. of the IEEE 13th International Conference on Computer Vision (ICCV 2011), Barcelona, Spain, November 2011.

LLD is a non-linear and non-parametric denoising algorithm for high-dimensional data with the intrinsic manifold strucutre. LLD denoises the manifold by jointly optimizing the local to global alignment error and graph Laplacian (Laplacian-Bertrami) energy.

 


 


Related Publications:

Dian Gong, Fei Sha and Gerard Medioni, "Locally Linear Denoising on Image Manifolds", Proc. of the 13th International Conference on Artifical Intelligence and Statistics (AISTATS 2010), Sardinia, Italy, May 2010. Journal of Machine Learning Research: W&CP 9.

Previous Works

  • Sparse Related Visual Learning, Self-taught Learning and Feature Extractions:

Though there are many image decomposition methods, it is hard to get both of the basis and the features to be independent without the normal distribution assumption. Recent research shows that sparseness and other constraints will lead to part-based representations results, which is similar to the receptive fields in V1 cortex in human Brain. Sparse Coding, Sparse Bayesian Learning and Compressive Sensing have been proposed for pattern learning, feature extraction, denoising and compression during the past 10 years.

 

In vision, self-taught learning means studying the knowledge from free-cost images in our natural environment; it is an active area in machine learning in recent years. The significance of self-taught learning is to revisit the fact that sometimes not only the labeled target data but also the relevant unlabeled data are hard to get, while at the same time the basic patterns can be embedded in the general data although it is unlabeled and with quite different distribution.

Propose a model-based feature extraction approach, which uses micro-structure modeling to design adaptive micro-patterns. We first model the micro-structure of the image by Pair-wise Markov Random Field. Then we give the generalized definition of micro-pattern based on the model. After that, we define the fitness function and compute the fitness index to encode the image’s local fitness to micro-patterns.

Papers: ICIP08, AMFG-ICCV05, US Patent     

  • Face Recognitions and Similarity Measurement:

Papers: ICIP05 and Microsoft Techfest Demo 2005

  • Geometric Methods and Applications:

Papers: Electronics Letters 2008 and CISS08

        

 

 

 

Publications

  • Journal

Dian Gong, Xuemei Zhao, Yunfan Li, "Tight geometric bound for Marcum Q-function",  IEE Electronic Letters, Volume 44, Feb. 28, 2008. [Zhao and Gong contributed equally to this work]

  • Conference

Dian Gong and Gerard Medioni, "Dynamic Manifold Warping for View Invariant Action Recognition", Proc. of the IEEE 13th International Conference on Computer Vision (ICCV 2011), Barcelona, Spain, November 2011. [Video] [Project Webpage] [new!]

Dian Gong, Fei Sha and Gerard Medioni, "Locally Linear Denoising on Image Manifolds", Proc. of the 13th International Conference on Artifical Intelligence and Statistics (AISTATS 2010), Sardinia, Italy, May 2010. Volume 9 of Journal of Machine Learning Research: W&CP 9. [Poster] [Project Webpage]

Dian Gong, Xuemei Zhao and Qiong Yang, "Sparse Non-Negative Pattern Learning for Image Represenation", Proc. of the 15th IEEE International Conference on Image Processing (ICIP 2008) ,San Diego, California, USA, October 2008. 

Qiong Yang, Dian Gong and Xiaoou Tang, "Modeling Micro-patterns for Feature Extraction",  Proc. of the 10th IEEE International Conference on Computer Vision (ICCV 2005), workshop on AMFG (Oral), pp.2-16, LCNS, Beijing, China, October 2005. 

Dian Gong, Qiong Yang, Xiaoou Tang, Jianhua Lu, “Extracting Micro-Structural Gabor Features for Face Recognition”, Proc. of the 12nd IEEE International Conference on Image Processing (ICIP 2005), Vol. 2, pp.924-5, Genova, Italy, September 2005.

Dian Gong, Yunfan Li, Xuemei Zhao, "Geometric Inversion Approach for Visual Curve Estimation", Proc. of the 42nd Annual Conference on Information Sciences and Systems (CISS 2008) (Oral), Princeton, NJ, USA, March 2008. 

Dian Gong, Zhiyao Ma, Yunfan Li, Wei Chen, Zhigang Cao, "High Order Geometric Range Free Localization in Opportunistic Cognitive Sensor Networks", Proc. of the 43rd IEEE International Conference on Communications (ICC 2008), CoCoNet (Oral), Beijing, China, 2008.

Dian Gong, Yunfan Li, "Dynamic System Analysis and Generalized Optimal Code Assignment of OVSF-CDMA Systems", Proc. of the 43rd IEEE International Conference on Communications (ICC 2008) (Oral), Beijing, China, 2008.

Dian Gong, Yusong Yan, Jianhua Lu, "Dynamic code assignment for OVSF code system ", Proc. of the 48th IEEE Global Telecommunications Conference (GLOBECOM 2005) (Oral), St. Louis, MO, USA, 2005.

  • Patents 

Qiong Yang, Dian Gong, Xiaoou Tang, Modeling Micro-Structure for Feature Extraction, US Patent 315262.02, 2007

Awards

Gold Medal, Contemporary Undergraduate Mathematical Contest in Modeling (CUMCM), 2002

International Mathematical Olympiad (IMO) Chinese National Team Candidate, 2000

Gold Medal, China Mathematics Olympiad (CMO), 2000

Miscellaneous

During the spare time, I love watching movies and listening pop music. Especially, I was a director in TDO (TDO means trade-off, it is extremely important for team work^_^) studio, Tsinghua University. We made two student movies and one music video. One of the movie is "how do I love you", which got the best idea and best actor prizes in the first Tsinghua Digital Movie Festival, 2004. 

Here is the link of this movie in youku. Actually, I just found it out by accident and I really do not know who put this online, but it is fine, enjoy it:-).


  

Links: Tie-Yan Liu, Jing ShengZhen Xiang, Chao Yu, Qiong Yang

Disclaimer

Disclaimer: IRIS group, Computer Science Department, USC Viterbi School of Engineering or University of Southern California do not screen or control the content on this website and thus does not guarantee the accuracy, integrity, or quality of such content. All content on this personal website is provided by and is the sole responsibility of the person from which such content originated..