Rachel Foster

General 9 min

Rachel Foster • 6 days ago

Token, AST, and Fingerprint Matching on Refactored Student Code

Renaming a variable, extracting a helper, and swapping a for loop for a while loop are the three moves students reach for when they want a copied submission to look original. Some similarity engines shrug them off, and some lose the match entirely. This piece walks through how token hashing, AST subtree matching, and fingerprinting each behave against a deliberately refactored Python pair, then compares what MOSS, JPlag, Dolos, and Codequiry actually reported on the same cohort.

Read

General 11 min

Rachel Foster • 1 week ago

How a Lecturer Catches Code Translated Between Languages

MOSS and JPlag compare Java to Java and Python to Python, which means a translated submission can score in single digits while the logic stays identical. This is how one lecturer, a TA, and a department chair handle ports, and what they've learned about the tooling that catches them.

Read

General 12 min

Rachel Foster • 1 week ago

How AST Comparison Catches Refactored Code Plagiarism

Renaming variables and swapping a for loop for a while loop defeats simple text matching, but it rarely defeats structural comparison. This report walks through the obfuscation ladder, the algorithms that climb it, and the published detection rates behind the claims, including the cases where every engine still misses.

Read

General 8 min

Rachel Foster • 3 weeks ago

The Evolution of Code Similarity Detection in One CS Department

A University of Missouri-St. Louis instructor spent two decades watching code similarity tools change shape. The gaps between each generation of detection explain why instructors now run peer, web, and AI checks side by side.

Read

General 8 min

Rachel Foster • 3 weeks ago

A TA's Script for Sorting 300 Code Similarity Reports by Office Hours

A teaching assistant at UC San Diego reduced a 312-submission similarity queue to a shortlist of 14 files in about two hours. The workflow relies on Codequiry's outlier scoring, a Python triage script, and a strict two-pass review rule. Here is the exact process, including the script and the thresholds she uses.

Read

General 11 min

Rachel Foster • 3 weeks ago

Benchmarking MOSS JPlag and Codequiry on 200 Student Submissions

A public university ran MOSS, JPlag, and Codequiry against the same 214 Python submissions, plus 30 AI-generated files. The tools disagreed on nearly a quarter of flagged cases. One combined approach changed how instructors review code.

Read

General 8 min

Rachel Foster • 1 month ago

AI Code Detector False Positives on Boilerplate

AI code detectors are producing false positives on the most ordinary submissions in CS1: code that looks the same because the assignment required it. This reported piece examines the data, the workflow changes instructors are making, and why combining AI detection with structural similarity reduces the error rate.

Read

General 6 min

Rachel Foster • 1 month ago

Automating Code Plagiarism Detection in GitHub Actions With Codequiry

Set up an automated code plagiarism detection pipeline in GitHub Actions using Codequiry's REST API. Follow precise steps to write a workflow YAML and a Python script that submits student code, receives similarity scores, flags suspicious pushes, and optionally detects AI-generated code. Includes threshold tuning, result interpretation, and false positive handling.

Read

General 7 min

Rachel Foster • 1 month ago

A Hiring Manager's Audit of AI-Generated Code in Take-Home Tests

An engineering leader audit of 1,284 remote take-home coding submissions found 31.2% flagged as likely AI-generated at high confidence. Manual review confirmed 279 of 401 high-confidence flags, with a 4.5% false positive rate among high-confidence flags. Here is what the data showed and how hiring managers should handle AI detection scores.

Read

General 10 min

Rachel Foster • 2 months ago

How Few AST Nodes Do You Need to Catch a Copied Function

A single function with renamed variables, reordered statements, and changed whitespace can still look structurally identical under the hood. This step-by-step guide builds a minimal AST clone detector in Python, explains where it breaks, and shows how production tools like Codequiry stack structural, token‑level, and web‑origin checks to catch the copying that student‑grade normalizers miss.

Read

General 7 min

Rachel Foster • 2 months ago

How UMass Amherst Brought AI Detection Into CS 121

When 800 students enroll in an introductory Python course, detecting AI-generated code by hand is impossible. UMass Amherst integrated an AI code detector trained on student-level patterns alongside traditional similarity checks—and uncovered a 14% AI flag rate, a 2% false positive rate, and a sustainable workflow that kept TAs focused on teaching rather than policing.

Read

General 7 min

Rachel Foster • 2 months ago

Across Two Semesters, AI Code Detector Accuracy Hit 87% in Python

A two-semester experiment at a mid-sized CS department put Codequiry’s AI code detector to the test across 1,200 student submissions. The tool achieved 87% overall accuracy in identifying AI-generated Python code, with a manageable false-positive rate and strong recall. The study surfaced distinct patterns in where detection excels—and where manual judgment remains essential.

Read