My Chinese name is Dong Ziqi, and I was born on September 13, 1997, China.
I am currently studying at Carnegie Mellon University, and my major is
My main academic direction is Software Engineering, I'm also a big fan of
I love music. I write lyrics and melody by myself and also compose & remix
I can play the guitar, bass, and midi-keyboard which is usually used in my
Logic Pro X is my favorite compose software, and I hope someday I can make
such excellent software.
I love design as well, not only photographs and pictures but also some web
pages. I use Affinity Photo as my main tool.
Designed a semantics-based mining algorithm that learns the code
patterns for library migrations and implemented a tool and automatically
generates the java code during migrations of library APIs.
Evaluated the effectiveness of three software testing approaches,
namely manual testing, monkey testing, and stochastic model-based
by counting the event-logs emitted by Android APK. Testing results
that Human Testing has the best performance in terms of UI events,
Testing has the best performance in terms of system events and works
indistinguishably differently from Human Testing with respect to event
coverage, Tool (Stoat) Testing has the best performance in terms of
lifecycle events and could mimic human behaviors for certain apps like
Amazon, Twitter, and Viber.
Developed the beta-version of an end-to-end smart logistic system that
is able to process operational command from mobile applications and
communicate with the core database for further business operations.
Implemented a deep-learning model using OpenCV and Caffe that is
capable to recognizing vehicle plates at the highway entrance to achieve
automatic payment through highway. Achieved 95%-97% accuracy in Neusoft
Conducting research in developing deep-learning-based models to assist
breast cancer screening and diagnosis, and established a deep-learning
model. We used TensorFlow's PNASnet to train breast cancer slice images
and achieved a final result with AUC = 0.9257 in a ROC test.
A neural network model is proposed to predict human pre-miRNA which is
based on Google's TensorFlow machine learning framework, so as to
achieve the aim of predicting miRNA. Responsible for data finding and
processing: Converting raw RNA sequence data to secondary structure.