Acemap: Academic Map System
- Develop visualizing applications for scholarly information networks and presentation approaches.
- Implement the recommending algorithm for papers in Acemap, and present the result on website.
- Build and maintain the server and the back-end for Acemap.
Academic Knowledge Graph System
- Cleaned and inferred on a huge academic dataset with billions of records including authors, papers, institutes and so on.
- Implement the Knowledge Graph System based on Jena framework consisting of TDB database and SPARQL engine, and present result and visualization on website.
- Implement network and triplet embedding algorithms including TransE, ComplEX, HolE and so on to predict and enrich records.
Large scale academic data visualization
- Visualize academic graphs with over 1 million nodes, layout of which is done using an N-body simulation. In order to simplify time complexity, Barnes-Hut algorithm is used based on quad-tree data structure.
- Modify source codes depending on Mysql C library, Cairo 2D graphics and Gtk+ 3.
Linux Kernel Applications Development
- Get understanding of how operating system works by developing Linux Kernel applications including adding vital system calls, dumping page table and completing page replacement algorithm.
- Score 100/100, Rank 1/143
Gene Chip Analysis
- Explore algorithms including deep forest, random forest, SVM and deep neural network with greedy layer-wise pre training on Gene-Classification task
- Score 96/100, Rank 2/84
Articulated Pose Estimation Using Convolutional Pose Machines
- Implements idea of paper ”Convolutional Pose Machines”, which designs CPMs consisting of a sequence of convolutional networks for the articulated pose estimation task.
- Score 97/100, Rank 3/47