English

Synthesizing Tasks for Block-based Programming

Computers and Society 2020-11-06 v3 Artificial Intelligence Machine Learning

Abstract

Block-based visual programming environments play a critical role in introducing computing concepts to K-12 students. One of the key pedagogical challenges in these environments is in designing new practice tasks for a student that match a desired level of difficulty and exercise specific programming concepts. In this paper, we formalize the problem of synthesizing visual programming tasks. In particular, given a reference visual task Tin\rm T^{in} and its solution code Cin\rm C^{in}, we propose a novel methodology to automatically generate a set {(Tout,Cout)}\{(\rm T^{out}, \rm C^{out})\} of new tasks along with solution codes such that tasks Tin\rm T^{in} and Tout\rm T^{out} are conceptually similar but visually dissimilar. Our methodology is based on the realization that the mapping from the space of visual tasks to their solution codes is highly discontinuous; hence, directly mutating reference task Tin\rm T^{in} to generate new tasks is futile. Our task synthesis algorithm operates by first mutating code Cin\rm C^{in} to obtain a set of codes {Cout}\{\rm C^{out}\}. Then, the algorithm performs symbolic execution over a code Cout\rm C^{out} to obtain a visual task Tout\rm T^{out}; this step uses the Monte Carlo Tree Search (MCTS) procedure to guide the search in the symbolic tree. We demonstrate the effectiveness of our algorithm through an extensive empirical evaluation and user study on reference tasks taken from the \emph{Hour of Code: Classic Maze} challenge by \emph{Code.org} and the \emph{Intro to Programming with Karel} course by \emph{CodeHS.com}.

Cite

@article{arxiv.2006.16913,
  title  = {Synthesizing Tasks for Block-based Programming},
  author = {Umair Z. Ahmed and Maria Christakis and Aleksandr Efremov and Nigel Fernandez and Ahana Ghosh and Abhik Roychoudhury and Adish Singla},
  journal= {arXiv preprint arXiv:2006.16913},
  year   = {2020}
}

Comments

NeurIPS 2020