About Me

Christy Yuan Li is a master student in Computational Data Science program at Carnegie Mellon University. During her master study, she has worked on various projects applying machine learning algorithms on real-world problems such as graph mining, natural language processing, video content retrieval, video captioning, and distributed system. Her research interests expand from machine learning to large scale data analysis on distributed networks. Before joining CMU, she earned her bachelor degree on Information Engineering from the Chinese University of Hong Kong. She has interned as a software engineer developing electronic signature web backend system.


Dec 2016

Carnegie Mellon University

Master in Computational Data Science

Coursework: Advanced Introduction to Machine Learning, Machine Learning with Large Datasets, Probability Graphical Model, Multimedia Database and Data Mining, Distributed System, Search Engine, Data Science Seminar, Convex Optimization, Topics in Deep Learning

July 2015

Chinese University of Hong Kong

Bachelor in Information Engineering



Energy-Constrained Distributed Learning and Classification by Exploiting Relative Relevance of Sensors’ Data

Majid Mahzoon, Christy (Yuan) Li, Xin Li and Pulkit Grover

in IEEE Journal on Selected Areas in Communications, vol. 34, no. 5, pp. 1417-1430, May 2016.


Energy efficient learning and classification for distributed sensing

Christy (Yuan) Li, X. Li and P. Grover

8th International Conference on Communication Systems and Networks (COMSNETS), Bangalore, 2016, pp. 1-6.


Novel Quantization Strategies for Linear Prediction with Guarantees

Simon Du, Yichong Xu, Christy (Yuan) Li, Hongyang Zhang, Pulkit Grover and Aarti Sign

International Conference on Machine Learning Workshop, June 2016


Constrained Ensemble Initialization for Facial Landmark Tracking in Video

Christy (Yuan) Li, Tadas Baltrusaitis, Louise-Phillipe Morency

submitted to 12th IEEE International Conference on Automatic Face and Gesture Recognition


Semantic Scene Search in Video

Christy (Yuan) Li*, Deepak Gopinath*, Jayanth Koushik*, Louise-Phillipe Morency

submitted to 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR)

Work Experience

May 2016 - Aug 2016

CMU Language Technologies Institute, School of Computer Science

Research Intern on Facial Landmark Tracking in Vide

Combined Project-Out Cascaded Regression with 3D shape model for detecting landmarks on images. Ex- tended the system to predict landmarks in video by exploiting temporal features.

May 2015 - Aug 2015

CMU Department of Electronic and Computer Engineering

Research Intern on Energy-Efficient Machine Learning on Distributed Sensor Network

Proposed a rate-allocation strategy and an adaptive learning algorithm for learning distribution-parameters of sensors under communication constraint. Applied the algorithms for Brain-Computer Interaction

June 2014 - Aug 2015

Applied Science Technology and Research Institution of Hong Kong

Research Intern on 2 view to multi-view conversion of 3D images

Implemented algorithms for generating 3D representations of images, including Block Motion Estimation and Compensation using SIFT features, and image warping by its depth map

Nov 2014 - Jan 2015

iASPEC, Hong Kong

Software Engineer

Develop E-Signature online deposit system. Hands-on experience in J2SE, J2EE, C3, Database, Web ser- vices, DAO, UML diagramming, Struts, Spring, Hibernate, HTML, etc.


Matlab, R, Python, C, C++, c♯, Java, Spark, Hive, Hadoop, TensorFlow, Theano, Keras, Torch, iPython, iTorch, Parameter Server, GraphLab, GraphX, GraphChi, Pig, GuineaPig, JavaScript, Node.js, HTML, WEKA, LISP, Prolog, XML, PHP, UML, MySQL.



Open Domain Video Description Alignment

Developped Multimodal Embedding model and End-to-End deep learning model to extract scene in movies or noisy clips that corresponds to given language descriptions, as well as learn compact representation of video content and language content. Created a noval dataset called Align-MVAD. Code writtern in Python and Theano.


MNIST Images Classification using Deep Neural Networks

Implemented Mixture of Bernoullis model, Neural Network, Restricted Boltzman Machine (RBM), Deep Restricted Boltzman Machine (DRBM), Autoencoder, Denoising Autoencoder, and Neural Networks with pertained RBM/Autoencoder initialization. Trained all models for classifying MNIST images and compared their performance. Code written in Python.


Character-Level Language Model using Recurrent Neural Network

Implemented Recurrent Neural Network (RNN) in Python for training character-level language model with tiny-shakespeare dataset.


Large-scale Document Classification

Implemented Streaming Naive Bayes algorithm using Java, Hadoop and GuineaPig respectively, and Gibbs Sampling LDA on Parameter Server for efficient training of large-scale document classification


Distributed SGD for Matrix Factorization on Spark

Implemented large-scale matrix factorization with distributed stochastic gradient descent in Spark for text modeling


Facial Expression Recognition

Developped a facial expression system that is robust for Asian and Western facial expressions, as well as created a novel large dataset on Asian expression. The system used random forest, SVM and deep neural network


Efficient Approximate PageRank and Visualization for Large-Scale Graph

Implemented snowball algorithm for a large-scale graph in Python and used Gephi visualization software to visualize the entire graph


Graph Mining using SQL

Extended graphMiner package with a k-core algorithm using Python and SQL. Conducted experiments on real world large graphs to discover graph patterns


Optimization of Deep Neural Networks Architecture by Bayesian Approach

Implemented a neural network for classification using Python. Proposed a threshold-based algorithm and randomized structural learning method to optimize neural network architecture


Distributed System

Implemented a distributed system supporting RPC between remote server and clients with robust scalability in real time environment by using caching, schedueling, network data analysis, and two-phase commit. Writtern in C and Java.


Search Engine

Implemented a search engine with boolean retrieval, ranked retrieval, strutured queries, query expansion, federated search, aggregated search, vertical search, selective search, and supervised learning algorithms for improving performance.

Contact Me

Address: 356 North Craig Street, Pittsburgh, 15213, PA

Call me : 4126090693

E-mail : liyuanchristy@gmail.com