UROP Proceedings 2020-21

School of Engineering Department of Computer Science and Engineering 143 Commonsense Reasoning with Knowledge Graphs Supervisor: SONG Yangqiu / CSE Student: ZHAO Jiachen / COMP Course: UROP1100, Spring Gamification is a method to motivate users by applying game elements to non-game situations. Building a satisfying gamified platform has advantages compared to ordinary ones, for example, more user activities and better outcomes. However, an unsuccessful gamification may have countereffects, and waste the efforts paid on it. In this study, we conduct a qualitative research to do sentiment analysis on the posts from Meta Stack Overflow. We will use a pre-trained natural language processing (NLP) model Bidirectional Encoder Representations from Transformers (BERT) to analyze the post texts to see whether gamification of Stack Overflow is satisfying and find out a pattern between questions and answers. Commonsense Reasoning with Knowledge Graphs Supervisor: SONG Yangqiu / CSE Student: WANG Weiqi / COSC Course: UROP1100, Summer Commonsense knowledge is crucial for artificial intelligence machines to understand natural language. While learning commonsense knowledge with the help of knowledge graph can improve the performance of deep learning models, most of the existing commonsense datasets use (ℎ, , ) formed CSV file as the dataset carrier, and the corresponding knowledge cannot be easily matched in the knowledge graph. In this project, we aim to align GLUCOSE, a commonsense dataset annotated by humans based on the ROCstory, to ASER, a large-scale eventuality knowledge graph, to investigate how much commonsense knowledge can be grounded in the eventuality knowledge graph. This resource can also serve as the preparation work for future studies on commonsense knowledge graph reasoning. Efficient Queries over Database Supervisor: WONG Raymond Chi Wing / CSE Student: ORAZALIN Alibek / MATH-CS Course: UROP1100, Fall Given two vertices s and t in a weighted graph, the shortest distance/path needs to be found efficiently where weights of edges may change over time. This query has many practical applications such as finding shortest path between two points on a map where time travelled for each part of the road may vary due to various reasons (traffic jams, weather conditions, etc.). Recent paper found an efficient way to find the shortest path in the graph in ( 4 ⋅ + ) time, while preprocessing takes ( ⋅ 2 ) and update takes only ( 3 ) time which is a significant improvement. This report presents the current progress regarding the research mentioned above.