We present TEGCER, an automated feedback tool for novice programmers. TEGCER uses supervised classification to match compilation errors in new code submissions with relevant pre-existing errors, submitted by other students before. The dense neural network used to perform this classification task is trained on 15,000+ error-repair code examples. The proposed model yields a test set classification Pred@3 accuracy of 97.7% across 212 error category labels. Using this model as its base, TEGCER presents students with the closest relevant examples of solutions for their specific error on demand. A large scale (N>230) usability study shows that students who use TEGCER are able to resolve errors more than 25% faster on average than students being assisted by human tutors.
Tue 12 NovDisplayed time zone: Tijuana, Baja California change
16:00 - 17:40 | Testing and VisualizationDemonstrations / Research Papers / Journal First Presentations at Cortez 1 Chair(s): Amin Alipour University of Houston | ||
16:00 20mTalk | History-Guided Configuration Diversification for Compiler Test-Program GenerationACM SIGSOFT Distinguished Paper Award Research Papers Junjie Chen Tianjin University, Guancheng Wang Peking University, Dan Hao Peking University, Yingfei Xiong Peking University, Hongyu Zhang The University of Newcastle, Lu Zhang Peking University | ||
16:20 20mTalk | Data-Driven Compiler Testing and Debugging Research Papers Junjie Chen Tianjin University | ||
16:40 20mTalk | Targeted Example Generation for Compilation Errors Research Papers Umair Z. Ahmed National University of Singapore, Renuka Sindhgatta Queensland University of Technology, Australia, Nisheeth Srivastava Indian Institute of Technology, Kanpur, Amey Karkare IIT Kanpur Link to publication Pre-print | ||
17:00 20mTalk | Lightweight Assessment of Test-Case Effectiveness using Source-Code-Quality Indicators Journal First Presentations Giovanni Grano University of Zurich, Fabio Palomba Department of Informatics, University of Zurich, Harald Gall University of Zurich Link to publication Pre-print | ||
17:20 10mDemonstration | Visual Analytics for Concurrent Java Executions Demonstrations Cyrille Artho KTH Royal Institute of Technology, Sweden, Monali Pande KTH Royal Institute of Technology, Qiyi Tang University of Oxford | ||
17:30 10mDemonstration | NeuralVis: Visualizing and Interpreting Deep Learning Models Demonstrations Xufan Zhang State Key Laboratory for Novel Software Technology Nanjing University, Nanjing, China, Ziyue Yin State Key Laboratory for Novel Software Technology Nanjing University, Nanjing, China, Yang Feng University of California, Irvine, Qingkai Shi Hong Kong University of Science and Technology, Jia Liu State Key Laboratory for Novel Software Technology Nanjing University, Nanjing, China, Zhenyu Chen Nanjing University | ||