Brian Harrington

Getting Creative with Semantic Similarity

This paper investigates how graph-based representations of entities and concepts can be used to infer semantic similarity and relatedness, and, more speculatively, how these can be used to infer novel associations as part of a creative process. We show how personalised PageRank on a co-occurrence graph can obtain competitive scores on a standard semantic similarity task, as well as being used to discover interesting and surprising links between entities. We hypothesise that such links could form the first stage in a creative ideation process.

TrAcademic: Experiences With Gamified Practical Sessions for a CS1 Course

This paper discusses preliminary impressions and results of using the TrAcademic system to gamify practical sessions in an introductory CS course at the University of Toronto Scarborough. Students were awarded points for attendance, completion of problems, and assisting other students in the session. Points were not counted towards the course grade, but were displayed on a public leaderboard. As a result of this simple change, practical attendance increased by over 500% from the previous offering, and both students and teaching assistants found the gamified practicals improved both engagement and retention.

Automatic detection of drug interaction mismatches in package inserts

The US Code of Federal Regulation (21 CFR 207) mandates that pharmaceutical manufacturers submit their FDA-approved drug information as medication Package Insert (PI). PI should provide comprehensive, current, and accurate information about the medical use of a drug. However, PI narratives are cumbersome for healthcare providers to navigate and are therefore rarely used by them. We are developing Paracelsus, a tool for automatically extracting structured drug information from PI. Paracelsus has the potential to allow healthcare providers to benefit from the rich drug information in PI for patient care. In this study, we report the development and evaluation of Paracelsus on drug-drug interactions. We show that Paracelsus performs with a high accuracy, discovering interactions not covered by other medical compendia, and in addition automatically detecting inconsistencies in the package inserts.

Tracing vs. Writing Code: Beyond the Learning Hierarchy

Much work has been done on the achievement gap between code tracing and code writing in CS1 students. The generally accepted explanation for this gap is that tracing and writing form separate steps in a learning scaffolding; students must first learn to trace code before they can be expected to write code. The expectation is that once students have mastered these skills, future grades will be driven by their ability to understand the deeper learning concepts, and so the gap between tracing and writing should disappear. In this paper, we detail and evaluate a study on 384 CS2 students to evaluate whether a tracing-writing gap still exists, and assess whether anything can be deduced about students who continue to exhibit such a gap. We find that not only does the gap seem to have closed by CS2, students are equally likely to show a reverse gap in the writing-tracing direction. However, further analysis shows a strong correlation between students who do continue to have a gap (in either direction) and poor overall performance in the course.

Get On Track, Stay On Track: Combining Early Intervention and Study Skills in a CS1 Seminar

This paper details our experiences running the OnTrack Seminar at the University of Toronto Scarborough. OnTrack is a one day seminar for CS1 students aimed at both stu- dents who are struggling, and those who are excelling and wish to ensure continued success. We found that holding this seminar in lieu of traditional early intervention methods improved participation and was better received by students across the grade spectrum. Early results indicate that this model may result in outcome improvements for students at risk in the early stages of CS1.

TrAcademic: Improving Participation and Engagement in CS1/CS2 with Gamified Practicals

Practice is an important part of introductory CS courses, and practical sessions are a students best opportunity for hands-on experience with the material covered in the course in a supervised, supportive environment. However, finding a balance between challenging more experienced students and alienating newcomers can be difficult and frustrating. One possible solution is to let the students self-select the problems they wish to attempt. The difficulty then becomes one of motivation and administration. This paper details our experiences with implementing a gamified system for practicals whereby students in CS1 and CS2 level courses receive points for various activities, including attendance, attempting basic problems, completing challenge problems and aiding fellow students with their work. These points have no bearing on final grades, but are displayed on a public leaderboard. In our experience, this gamified system dramatically improved attendance at practical sessions, was well received by students and TAs alike, improved retention rates, and offered students the opportunity and motivation to overcome poor performance early in the course and improve their results.

ICSC 2013: Message from the program chairs

Welcome to the proceedings of the Seventh IEEE International Conference on Semantic Computing (ICSC 2013) in Irvine, California, USA. ICSC 2013 continues to foster the growth of a new research community. The conference builds on the tradition and success of the past ICSC series as an international forum for researchers and practitioners in academia and industry to present research that advances the state of the art and practice of Semantic Computing, identifies the emerging research topics, and defines the future of the field.