School of Engineering Department of Computer Science and Engineering 136 Handling User Challenge in Human-Agent Interaction Supervisor: MA Xiaojuan / CSE Student: LIU Yuzhi / MATH-AM Course: UROP2100, Summer It is necessary for researchers to have the ability to read papers effectively throughout their academic lives. However, this can be difficult for researchers who are new to their fields. For instance, new researchers may not be able to question the papers and find the appropriate answers, which is a critical process in paper reading. Human-computer interaction (HCI) researchers have explored several ways to facilitate the reading experience. Based on the previous methods, we propose an interactive agent that can generate questions automatically based on the users’ browsing behavior. Machine Learning Approach to Creating Personalized E-Learning Paths Supervisor: MA Xiaojuan / CSE Student: GUO Bingcan / DSCT Course: UROP1100, Fall Community Question Answering (CQA) platforms have been an important source of knowledge sharing. Users can raise(answer) questions and rate other’s questions(answers) on the websites. By doing so, they will receive reputation points and badges to signify their contribution. However, while quality questions attract constant discussion, some other questions, even with similar content, are ignored. Hence, we want to find what makes an effective question that receives answers soon, scores high or somehow contributes to the CQA community. By carrying out quantitative data analysis, we hope to uncover traits of effective questions and propose possible guidelines for asking questions more efficiently. In this report, data extracted from the Stack Overflowwebsite are studied. We summarize statistical features in table format and carry out hypothesis testing. Preliminary findings and plans for the next phase will be discussed eventually. Machine Learning Approach to Creating Personalized E-Learning Paths Supervisor: MA Xiaojuan / CSE Student: ZHOU Yuran / COGBM Course: UROP2100, Fall The stack overflow platform has become the CQA platform with the highest traffics and attention. After researching and extracting data from the top three computational and non-computational CQAs from the previous project, a decision is made to focus more on Stack Overflow from more detailed features and further implications. With new tools and methods implemented, we are now able to mine and extrapolate more information from this website to perform further analysis. As a part of a larger research, this project will mainly focus on the implication of the number of verbs within the question title and how it may affect the quality of the question via multiple standards. In this report, the means of data extraction and manipulation will be explained, the specific implication of verbs will be discussed, and the final recommendation of wording will be given to facilitate CQA users to raise popular questions.