We discovered that understanding physical causality brings verb acquisition and noun grounding into a bootstrapping loop with 5.2% to 13.3% improvement, which is significant under McNemar's test and t-test.
We proposed a novel FF-GNN to improve graph fingerprinting methods by parameterizing with a multi-channel fuzzy histogram, which matches the state-of-the-art performance and has high computational efficiency.
Implemented the PageRank algorithm with Scala and Spark to find the most influential users in a Twitter social graph, and deployed the Spark applications on Azure HDInsight and Databricks.
Implemented a C++ version of the game 2048 which responded to player's keystrokes, and added support for customized tiles such as Unicode Emojis.
We used MATLAB to solve a nonlinear optimization task from scratch, aiming to find the minimum for an objective function which integrates the exponential of polynomial function with 101 degrees.
Based on "Breast Cancer Wisconsin Data Set", applied principal component analysis to explore the combination of most informative predictors, and compared different methods like QDA and random forest on R.