Identifying AI Generated Code with Parallel KNN Weight Outlier Detection
Lecture Notes in Networks and Systems, ISSN: 2367-3389, Vol: 1140, Page: 459-470
2025
Metric Options: CountsSelecting the 1-year or 3-year option will change the metrics count to percentiles, illustrating how an article or review compares to other articles or reviews within the selected time period in the same journal. Selecting the 1-year option compares the metrics against other articles/reviews that were also published in the same calendar year. Selecting the 3-year option compares the metrics against other articles/reviews that were also published in the same calendar year plus the two years prior.
Example: if you select the 1-year option for an article published in 2019 and a metric category shows 90%, that means that the article or review is performing better than 90% of the other articles/reviews published in that journal in 2019. If you select the 3-year option for the same article published in 2019 and the metric category shows 90%, that means that the article or review is performing better than 90% of the other articles/reviews published in that journal in 2019, 2018 and 2017.
Citation Benchmarking is provided by Scopus and SciVal and is different from the metrics context provided by PlumX Metrics.
Example: if you select the 1-year option for an article published in 2019 and a metric category shows 90%, that means that the article or review is performing better than 90% of the other articles/reviews published in that journal in 2019. If you select the 3-year option for the same article published in 2019 and the metric category shows 90%, that means that the article or review is performing better than 90% of the other articles/reviews published in that journal in 2019, 2018 and 2017.
Citation Benchmarking is provided by Scopus and SciVal and is different from the metrics context provided by PlumX Metrics.
Book Chapter Description
Plagiarism is an emerging issue in programming, and it becomes more difficult to identify due to generative AI; it can easily provide solutions on assignments that are supposed to be completed without any help. Many AI detectors are only focused on general text. We present an AI detector for programming based on KNN weight outlier detection. On assignments disallowing AI help, only few students still insist to use AI. Further, AI generated code is different to that of undergraduates. The detector employs six intermediate representations: text strings, token strings, generalized token strings, expanded token strings, linearized syntax trees, and linearized parse trees. According to our experiment on three data sets with MAP and processing time as the performance metrics, our detector is satisfactorily effective and efficient. The most optimal performance occurs when the intermediate representation is the token strings, the n for n-grams is 110, and the number of clusters is 3.
Bibliographic Details
http://www.scopus.com/inward/record.url?partnerID=HzOxMe3b&scp=85212485668&origin=inward; http://dx.doi.org/10.1007/978-3-031-71530-3_29; https://link.springer.com/10.1007/978-3-031-71530-3_29; https://dx.doi.org/10.1007/978-3-031-71530-3_29; https://link.springer.com/chapter/10.1007/978-3-031-71530-3_29
Springer Science and Business Media LLC
Provide Feedback
Have ideas for a new metric? Would you like to see something else here?Let us know