@inproceedings{ramesh-wu-2026-team,
title = "Team Yuvan at {S}em{E}val-2026 Task 13: Task-Adaptive Ensemble Strategies for {AI}-Generated Code Detection",
author = "Ramesh, Yuvan and
Wu, Tongtong",
editor = "Kochmar, Ekaterina and
Ghosh, Debanjan and
North, Kai and
Komachi, Mamoru",
booktitle = "Proceedings of the 20th {I}nternational {W}orkshop on {S}emantic {E}valuation (2026)",
month = jul,
year = "2026",
address = "San Diego, California, USA",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2026.semeval-1.263/",
pages = "2089--2097",
ISBN = "979-8-89176-414-9",
abstract = "We describe our system for SemEval-2026 Task 13 on detecting machine-generated code across eight programming languages and three subtasks: binary human-vs-AI detection, 11-way source identification, and 4-way generator classification. Our approach uses a task-specific combination of Qwen2.5-Coder-1.5B with LoRA fine-tuning, abstract syntax tree (AST) features, CodeBERT with head-tail chunking, and TF-IDF features. Experiments reveal three main findings. For Task A, neural detectors degrade markedly on the official test split, while AST-based structural features remain more stable, suggesting substantial distribution shift. For Task B, inverse-frequency class weighting is essential under extreme label imbalance and greatly improves macro-F1. For Task C, combining neural and statistical models performs better than relying on a single model alone, indicating complementary strengths across representations. Our final system achieves 0.638 macro-F1 on Task A, 0.449 macro-F1 on Task B, and 0.714 macro-F1 on Task C, offering practical insights into robustness, imbalance handling, and model complementarity for AI-generated code detection."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="ramesh-wu-2026-team">
<titleInfo>
<title>Team Yuvan at SemEval-2026 Task 13: Task-Adaptive Ensemble Strategies for AI-Generated Code Detection</title>
</titleInfo>
<name type="personal">
<namePart type="given">Yuvan</namePart>
<namePart type="family">Ramesh</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Tongtong</namePart>
<namePart type="family">Wu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2026-07</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Proceedings of the 20th International Workshop on Semantic Evaluation (2026)</title>
</titleInfo>
<name type="personal">
<namePart type="given">Ekaterina</namePart>
<namePart type="family">Kochmar</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Debanjan</namePart>
<namePart type="family">Ghosh</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Kai</namePart>
<namePart type="family">North</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Mamoru</namePart>
<namePart type="family">Komachi</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>Association for Computational Linguistics</publisher>
<place>
<placeTerm type="text">San Diego, California, USA</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
<identifier type="isbn">979-8-89176-414-9</identifier>
</relatedItem>
<abstract>We describe our system for SemEval-2026 Task 13 on detecting machine-generated code across eight programming languages and three subtasks: binary human-vs-AI detection, 11-way source identification, and 4-way generator classification. Our approach uses a task-specific combination of Qwen2.5-Coder-1.5B with LoRA fine-tuning, abstract syntax tree (AST) features, CodeBERT with head-tail chunking, and TF-IDF features. Experiments reveal three main findings. For Task A, neural detectors degrade markedly on the official test split, while AST-based structural features remain more stable, suggesting substantial distribution shift. For Task B, inverse-frequency class weighting is essential under extreme label imbalance and greatly improves macro-F1. For Task C, combining neural and statistical models performs better than relying on a single model alone, indicating complementary strengths across representations. Our final system achieves 0.638 macro-F1 on Task A, 0.449 macro-F1 on Task B, and 0.714 macro-F1 on Task C, offering practical insights into robustness, imbalance handling, and model complementarity for AI-generated code detection.</abstract>
<identifier type="citekey">ramesh-wu-2026-team</identifier>
<location>
<url>https://aclanthology.org/2026.semeval-1.263/</url>
</location>
<part>
<date>2026-07</date>
<extent unit="page">
<start>2089</start>
<end>2097</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T Team Yuvan at SemEval-2026 Task 13: Task-Adaptive Ensemble Strategies for AI-Generated Code Detection
%A Ramesh, Yuvan
%A Wu, Tongtong
%Y Kochmar, Ekaterina
%Y Ghosh, Debanjan
%Y North, Kai
%Y Komachi, Mamoru
%S Proceedings of the 20th International Workshop on Semantic Evaluation (2026)
%D 2026
%8 July
%I Association for Computational Linguistics
%C San Diego, California, USA
%@ 979-8-89176-414-9
%F ramesh-wu-2026-team
%X We describe our system for SemEval-2026 Task 13 on detecting machine-generated code across eight programming languages and three subtasks: binary human-vs-AI detection, 11-way source identification, and 4-way generator classification. Our approach uses a task-specific combination of Qwen2.5-Coder-1.5B with LoRA fine-tuning, abstract syntax tree (AST) features, CodeBERT with head-tail chunking, and TF-IDF features. Experiments reveal three main findings. For Task A, neural detectors degrade markedly on the official test split, while AST-based structural features remain more stable, suggesting substantial distribution shift. For Task B, inverse-frequency class weighting is essential under extreme label imbalance and greatly improves macro-F1. For Task C, combining neural and statistical models performs better than relying on a single model alone, indicating complementary strengths across representations. Our final system achieves 0.638 macro-F1 on Task A, 0.449 macro-F1 on Task B, and 0.714 macro-F1 on Task C, offering practical insights into robustness, imbalance handling, and model complementarity for AI-generated code detection.
%U https://aclanthology.org/2026.semeval-1.263/
%P 2089-2097
Markdown (Informal)
[Team Yuvan at SemEval-2026 Task 13: Task-Adaptive Ensemble Strategies for AI-Generated Code Detection](https://aclanthology.org/2026.semeval-1.263/) (Ramesh & Wu, SemEval 2026)
ACL