Chao Qin

Address: **** *. **** ****, ******, CA *5343 Tel: 209-***-****

WWW: https://eng.ucmerced.edu/people/cqin E-mail: abqso1@r.postjobfree.com

Objective

Seeking a software engineer in a challenging and active team

Summary of Qualifications

In-depth understanding of topics in machine learning and speech technology, data mining, text processing,

Extensive hands-on experience in developing and applying machine learning techniques to speech applications

Pro cient in algorithm development and software implementation

Excellent coding skills in C/C++, Matlab and scripting in Unix/Linux environment

Highly independent and self-motivated

Scholarships and Awards

Sep. 2007 to Mar. 2008: Marie Curie Fellowship for Early Stage Researcher from European Union

Aug. 2007: Best student paper award (Interspeech 2007) by International Speech Communication Association

Education

Aug. 2007 to May. 2011: PhD in Computer Science, University of California, Merced, USA (GPA: 4/4)

Aug. 2003 to Aug. 2005: MPhil in Electronic Engineering, The Chinese University of Hong Kong, PRC (GPA: 3.8/4)

Sep. 1999 to Jun. 2003: BSc in Electronic Information, Nanjing University, PRC (GPA: 3.6/4)

Technical Expertise

Knowledge

Machine learning: supervised/unsupervised learning, mixture models, neural networks, sequential data, latent variables

Speech: articulatory/acoustic modeling, speech recognition/synthesis, speaker adaptation/recognition, speech animation

Data mining: classi cation, regression, clustering, data analysis, web mining

Large-scale computing: MapReduce and Hadoop (basic)

Computing

Languages: Matlab and C/C++ (pro cient), Java (prior experience), Python (basic), Perl and Bash (intermediate)

Operation systems: Unix/Linux (pro cient), Mac (prior experience)

Tools: HTK and Edinburgh Speech Tools for speech processing (pro cient), SVN for version control (prior experience)

Research Experience

Mar. 2008 to May. 2011: research assistant at EECS, School of Engineering, University of California, Merced

Working on data-driven approaches to articulatory speech processing

Developed algorithms to adapt predictive models of tongue by a ne transformation and numerical optimization

Developed algorithms for missing data reconstruction by probabilistic mixture model and unsupervised learning

Developed software for data pre-processing and visualization of large-scale articulatory speech databases

Advisor: Prof. Miguel A. Carreira-Perpi n

na

Sep. 2007 to Mar. 2008: research fellow at The Centre for Speech Technology Research, University of Edinburgh

Worked on predictive modeling of tongue shapes

Developed landmark based predictive models for tongue by multivariate linear regression and neural networks

Advisors: Prof. Steve Renals and Dr. Korin Richmond

Sep. 2005 to Jul. 2007: research assistant at Adaptive System Lab, OGI/Oregon Health and Science University

Worked on data-driven approaches to articulatory inversion and robotic arm inverse kinematics

Developed algorithms for articulatory inversion by regression, mean-shift, mixture model, probabilistic tracking

Implemented software for Gaussian mixture model, Gaussian mean-shift, and feature extraction from speech

Advisor: Prof. Miguel A. Carreira-Perpi n

na

Aug. 2003 to Aug. 2005: research assistant at DSP and Speech Lab, The Chinese University of Hong Kong

Worked on automatic speech recognition and speaker recognition

Developed speaker authentication by combining speaker and utterance veri cation via support vector machine

Implemented hidden Markov model based Cantonese digit recognition and MFCC acoustic front-end in C

Advisor: Prof. Tan Lee and Prof. Pak-Chung Ching

Publications

C. Qin and M. A. Carreira-Perpi n (2010): Articulatory inversion of American English / / by conditional density

na

modes. Proc. Interspeech 2010 (talk), pp. 1998 2001.

C. Qin and M. A. Carreira-Perpi n and Mohsen Farhadloo (2010): Adaptation of a tongue shape model by local

na

feature transformations. Proc. Interspeech 2010 (poster), pp. 1596 1599.

C. Qin and M. A. Carreira-Perpi n (2010): Estimating missing data sequences in X-ray microbeam recordings. Proc.

na

Interspeech 2010 (poster), 1592 1595.

N. H. Zheng, C. Qin, Tan Lee, and P. C. Ching (2010): CU2C: A dual-condition Cantonese speech database for speaker

recognition applications. Book project, to appear.

C. Qin and M. A. Carreira-Perpi n (2010): Reconstructing the full tongue contour from EMA/X-Ray microbeam.

na

IEEE Int. Conf. on Acoustics, Speech and Signal Processing (ICASSP 2010) (talk), pp. 4190 4193.

C. Qin and M. A. Carreira-Perpi n (2009): The geometry of the articulator region that produces a speech sound.

na

43th Annual Asilomar Conference on Signals, Systems, and Computers (ASILOMARSSC 2009) (talk), pp. 1742 1746.

C. Qin and M. A. Carreira-Perpi n (2009): Adaptation of a predictive model of tongue shapes. Proc. Interspeech

na

2009 (poster), pp. 772 775.

C. Qin, M. A. Carreira-Perpi n, K. Richmond, A. Wrench, and S. Renals (2008): Predicting tongue shapes from a

na

few landmark locations. Proc. Interspeech 2008 (talk), pp. 2306 2309.

C. Qin, M. A. Carreira-Perpi n (2008): Trajectory inverse kinematics by conditional density modes. IEEE Int.

na

Conf. on Robotics and Automation (ICRA 2008) (talk), pp. 1979 1986.

C. Qin and M. A. Carreira-Perpi n (2008): Trajectory inverse kinematics by nonlinear, nongaussian tracking. IEEE

na

Int. Conf. on Acoustics, Speech and Signal Processing (ICASSP 2008) (poster), pp. 2057-2060.

C. Qin and M. A. Carreira-Perpi n (2007): An empirical investigation of the nonuniqueness in the acoustic-to-

na

articulatory mapping. Proc. Interspeech 2007 (talk), pp. 74 77. Best student paper award.

C. Qin and M. A. Carreira-Perpi n (2007): A comparison of acoustic features for articulatory inversion. Proc.

na

Interspeech 2007 (talk), pp. 2469 2472.

C. Qin, Tan Lee, and H. Meng (2005): On anti-model design for Cantonese verbal information veri cation. Proc. 8th

National Conference on Man-Machine Speech Communication (NCMMSC 2005) (talk).

N. H. Zheng, C. Qin, Tan Lee, and P. C. Ching (2005): CU2C: A dual-condition Cantonese speech database for speaker

recognition applications. Proc. Oriental COCOSDA (COCOSDA 2005) (talk), pp. 67 72.

C. Qin and Tan Lee (2004): Cantonese verbal information veri cation system using GMM-based anti-models. Proc.

Int. Sym. Chinese Spoken Language Processing (ISCSLP 2004) (poster), pp. 297 300.

Professional Activities

Member of: A liated member of IEEE Signal Processing Society, IEEE student member

Reviewer: Journal of Acoustical Society of America Express Letter (2010), IEEE Transaction on Robotics (2009),

IEEE Conference on Automation Science and Engineering (CASE 2008), IEEE International Conference on Robotics

and Biomimetics (ROBIO 2008), Neurocomputing (2007)

Language

English ( uent), Mandarin (native tongue), Cantonese ( uent)

Contact this candidate