Name
Abstract | ||
|---|---|---|
Computing theory is an important part of computer science education, introducing students to computational models of increasing power to study possibilities and limitations of computation. The subject is, however, very abstract and mathematical, and students often struggle with it. Students must master various computational models, but there is often a lengthy delay from the time a model is introduced until a student gets feedback on their related assignment. During this time, the course has typically moved far ahead, and students become progressively more lost. To alleviate this problem, we developed a prototype of an automated feedback tool for CS theory, which extends the widely used JFLAP software. Our tool currently handles student submissions of deterministic and non-deterministic finite automata, regular expressions, context-free grammars, and push-down automata homework, where an instructor specifies the target language and the students receive immediate feedback on their submissions. Currently, for incorrect submissions, the feedback is in the form of a "witness'' string, specifying a string on which the submission fails. Beyond regular languages, our tool attempts to solve undecidable problems; fortunately, the undecidability does not occur on typical homework assignments. We are collecting preliminary evaluation data from students using the prototype tool in their course. In our future work, we will analyze the data, and we aim to produce automated partial credit (along with the witness feedback) using SAT and QBF solvers.
|
Year | DOI | Venue |
|---|---|---|
2020 | 10.1145/3328778.3372598 | SIGCSE |
DocType | ISBN | Citations |
Conference | 978-1-4503-6793-6 | 0 |
PageRank | References | Authors |
0.34 | 0 | 5 |
Name | Order | Citations | PageRank |
|---|---|---|---|
| Ivona Bezáková | 1 | 0 | 3.72 |
| Edith Hemaspaandra | 2 | 1341 | 86.98 |
| Aryeh Lieberman | 3 | 0 | 0.34 |
| Hannah Miller | 4 | 0 | 0.68 |
| David E. Narváez | 5 | 0 | 3.04 |