This is a trailor of SHERLOCK - my machine learning final project. I’ve really been obsessively perfecting my final course thesis (as we’ve talked about it before here). Though immensely energy and time consuming, they’re just for fun but I’m proud of them for their beauty.
and the result is pretty nice as well.
Abstract
We introduce SHERLOCK, an innovative machine learning model designed to differentiate human-written code from AI-generated code in educational settings. By combining perturbation analysis with advanced machine learning, SHERLOCK significantly outperforms existing detection techniques, achieving a precision rate over 0.8 and a model score beyond 0.75. Our contribution addresses a critical need in maintaining academic integrity in programming courses amidst the rise of AI coding tools, marking an interesting direction to the field of AI-generated content detection.
The course thesis is long… cause students are incentivized towards GPA after all. But here we might be more genuine and, the plan is, to write a short, clear explain of this project. Stay tuned for tmrw!