Yuntao Lu

UT-Austin | USTC | UESTC

About me





Yuntao (Vicky) Lu

I am an optimistic and reliable person with love to explore new fantastic skills and technologies. And I would like to work in a team and corporate with other excellent people. In my graduate and undergraduate degrees, I learnt Machine Learning related courses and implemented various projects, which made me interested in this field and hope to acquire more knowledge in the future work. :)

More about me

Follow me

Education

University of Texas at Austin 2018 - 2020
School of Information | Information Studies | CPA 3.76

University of Science and Technology of China 2015 - 2018
School of Computer Science and Technology | Computer Systems and Architectures | GPA 3.21

University of Electronic Science and Technology of China 2011 - 2015
School of Information and Software Engineering | Embedded Systems | GPA 3.84

Projects

Profiling Information of Environmental Nonprofit Organizations in Texas Available Link
• Retrieved and processed data from AWS and opened tax files datasets with Python Pandas, NLTK, and IRSx packages.
• Trained a 5-dense-layer neural network classifier with TensorFlow and classified taxpayers into 26 domain-related groups.
• Retrieved and constructed a panel dataset of taxpayers across five years, in terms of environmental taxpayers of Texas.

An Application to Visualize the Partial World Indicator Data. Available Link
• Retrieved and grouped life expectancy and air pullution feaures from the World Indicator Dataset by countries.
• Designed and deployed an reactive application on a server to display linear and geographic figures by Python Plotly and Dash packages.

Analysis of the IMDB Movie Data with Machine Learning Algorithms.
• Preprocessed IMDB movie data by cleaning and encoding text features with Python Pandas and NLTK packages.
• Trained a 5-dense-layer neural network classifier with TensorFlow and classified taxpayers into 26 domain-related groups.
• Conducted the grid search of hyperparameters for classifiers, and evaluated with the accuracy, precision, recall and hmean metrics.

Text Detection in Natural Scene Images with Convolutional Neural Networks Available Link
• Trained a convolutional neural network and generate coordinates of bounding boxes for texts on datasets of International Conference on Document Analysis and Recognition (TensorFlow on a Microsoft Azure two-GPU virtual machine).
• Trained a 5-dense-layer neural network classifier with TensorFlow and classified taxpayers into 26 domain-related groups.
• Evaluated and shrank geometry coefficients to the ground truth with precision, recall and hmean metrics.

An FPGA Accelerator for Sparse Neural Networks Available Link
• Extracted parameters of popular neural networks with Caffe deep learning framework, and compressed sparse weight parameters with COO/CSR compression method.
• Designed architecture and computation functions for the prediction stage of sparse neural networks.
• Optimized accelerating cores with pipelines and buffers, and simulated computing cycles with Xilinx Vivado HLS.

Articles

Skills





Programming Languages

C/C++, Python, R, SQL

Software Tools

Jupyter Notebook, Pycharm, Chameleon Cloud Platform, AWS, Microsoft Azure, R Studio

Systems

Mac OS, Linux (Ubuntu), Windows

Machine Learning / Data Science

Keras, Scikit-Learn, NLTK, Pandas, Plotly

Contact

+8618550808418 | +1(512)-586-1553

luyuntao@mail.ustc.edu.cn | yuntaolu@utexas.edu

Github https://github.com/yuntaolu