JWare Lab

📣 ICSE 2027 Competition · Dublin, Ireland

The Clarification Challenge — clarify before you code

Can your coding assistant ask the right question? We are co-organizing the Clarification Challenge at ICSE 2027: build an assistant that spots underspecified requirements, asks the right clarifying question, and then writes code that matches what the user actually meant. The winning team gets free ICSE registration, and every ranked entry may submit a solution paper to the ICSE 2027 Competition Track.

Open now: enter by pull request · Solution papers due Dec 4, 2026 · Final submission date to be announced

Organized by researchers at the University of Luxembourg and Michigan Tech.

Join the competition →

JWare Lab (杰件实验室) is affiliated with the Department of Computer Science at Michigan Technological University. The name JWare is inspired by the PI’s first name, Jie (pronounced “jyeh”), a character meaning greatness. This reflects our aspiration: that software, built with the right foundations, can be truly great. We also aspire to empower students with the foundations to build great software in their future careers — in an era increasingly shaped by Generative AI.

We investigate the quality foundations, principles, and methodologies that make AI-assisted software engineering rigorous and trustworthy. Our research spans the full software development lifecycle: design, coding, testing, review, deployment, and maintenance. We are committed to ensuring that AI tools elevate the quality and safety of software, rather than erode it.

Lab Members

Faculty

Dr. Jie Wu Principal Investigator

PhD Students

Joshana Shakya PhD Student (started Fall 2026)

Indrajeet Roy PhD Student (started Fall 2026)

Josh Dafoe PhD Student (co-advised by Prof. Bo Chen)

Undergraduate Students

Gabe Dautovi Undergraduate Researcher · URS Scholar

Andrew Sadler Undergraduate Researcher · CC-REU Fellow

News

2026 — Ali Afoud’s first-author paper, “Verification as an Architectural Layer for LLM Agents”, is accepted at the JAW 2 Workshop at ASE 2026. Congrats, Ali!
Summer 2026 — Gabe Dautovi is named a Michigan Tech Undergraduate Research Scholar (URS). Congrats, Gabe!
Summer 2026 — Andrew Sadler is selected as a CC-REU Fellow. Congrats, Andrew!
Jun 2026 — Indrajeet Roy and Joshana Shakya join as incoming PhD students (Fall 2026)
Jun 2026 — JWare Lab website was launched
Jun 2026 — Research positions open for undergrad and graduate students — Summer & Fall 2026

Join Us

🎓 PhD Positions: I am actively recruiting highly-motivated students to join my research group at Michigan Tech, a Carnegie R1 (Very High Research Activity) institution, to work at the intersection of Software Engineering and AI, particularly Large Language Models. Admission is competitive: applicants need to be accepted by the graduate school as top candidates. See more details in this link. I apologize that I’m unable to reply to each email individually. A practical tip: the strongest applicants apply directly to the PhD program rather than waiting for my reply, and we talk once they stand out in the applicant pool.

📧 Research Assistant Positions at MTU: Paid undergraduate and graduate research opportunities are available on a rolling basis in Agentic Software Engineering for MTU students. If you are interested, please see this link.

See our open positions for full details on both.

Preprint

VeeReview

VeeReview: Measuring Build–Verify Pairs in Agent-Authored Pull Requests from a V-Model Perspective

Gabriel Dautovi, Jie JW Wu

Arxiv Code @article{dautovi2026veereview, title={VeeReview: Measuring Build--Verify Pairs in Agent-Authored Pull Requests from a V-Model Perspective}, author={Dautovi, Gabriel and Wu, Jie JW}, journal={arXiv preprint arXiv:2610.29143}, year={2026} }

(Under Review)
AssumptionMiner

AssumptionMiner: Extracting, Tracing, and Revising Implicit Assumptions in LLM Code Generation

Jie JW Wu

Arxiv Code Blog @article{wu2026assumptionminer, title={AssumptionMiner: Extracting, Tracing, and Revising Implicit Assumptions in LLM Code Generation}, author={Wu, Jie JW}, journal={arXiv preprint arXiv:2607.22898}, year={2026} }

(Under Review)
Overthink Attacks

Overthink-Triggered Slowdown Attacks on LVLM-Based Robotic Systems

Qiang Han, Jie JW Wu, Bo Chen

Arxiv @article{han2026overthink, title={Overthink-Triggered Slowdown Attacks on LVLM-Based Robotic Systems}, author={Han, Qiang and Wu, Jie and Chen, Bo}, journal={arXiv preprint arXiv:2607.01518}, year={2026} }

arXiv preprint, 2026
MANTRA

MANTRA: a Framework for Multi-stage Adaptive Noise TReAtment During Training

Shawn Zhao, Fatemeh H. Fard, Jie JW Wu

ArXiv @article{zhao2025mantra, title={MANTRA: a Framework for Multi-stage Adaptive Noise TReAtment During Training}, author={Zhao, Shawn and Fard, Fatemeh H. and Wu, Jie JW}, journal={arXiv preprint arXiv:2512.04319}, year={2025} }

(Under Review)
ClarifyCoder

Can Code Language Models Learn Clarification-Seeking Behaviors?

Jie JW Wu, Manav Chaudhary, Davit Abrahamyan, Arhaan Khaku, Anjiang Wei, Fatemeh H. Fard

Arxiv Code Video @article{wu2025clarifycoder, title={Can Code Language Models Learn Clarification-Seeking Behaviors?}, author={Wu, Jie JW and Chaudhary, Manav and Abrahamyan, Davit and Khaku, Arhaan and Wei, Anjiang and Fard, Fatemeh H.}, journal={arXiv preprint arXiv:2504.16331}, year={2025} }

(under review)

Publications

2027

Clarification Challenge

The Clarification Challenge: Towards Coding Assistants that Clarify Unclear User Requirements

Amal Akli, Jie JW Wu, Cedric Richter, Mike Papadakis

Competition @inproceedings{akli2027clarification, title={The Clarification Challenge: Towards Coding Assistants that Clarify Unclear User Requirements}, author={Akli, Amal and Wu, Jie JW and Richter, Cedric and Papadakis, Mike}, booktitle={Proceedings of the IEEE/ACM 49th International Conference on Software Engineering (ICSE), Competition Track}, year={2027} }

IEEE/ACM 49th International Conference on Software Engineering (ICSE 2027), Competition Track

2026

V-Model for Agents

Verification as an Architectural Layer for LLM Agents: A V-Model Design, and a Pilot Study of Its Deterministic Core

Ali Afoud, Jie JW Wu

Arxiv @article{afoud2026verification, title={Verification as an Architectural Layer for LLM Agents: A V-Model Design, and a Pilot Study of Its Deterministic Core}, author={Afoud, Ali and Wu, Jie JW}, journal={arXiv preprint arXiv:2609.31937}, year={2026} }

JAW 2 Workshop at the 41st IEEE/ACM International Conference on Automated Software Engineering (ASE 2026)
MergeRepair

MergeRepair: An Exploratory Study on Merging Task-Specific Adapters in Code LLMs for Automated Program Repair

Meghdad Dehghan, Jie JW Wu, Fatemeh Hendijani Fard, Ali Ouni

Paper Arxiv Code @article{dehghan2024mergerepair, title={MergeRepair: An Exploratory Study on Merging Task-Specific Adapters in Code LLMs for Automated Program Repair}, author={Dehghan, Meghdad and Wu, Jie JW and Fard, Fatemeh H and Ouni, Ali}, journal={Empirical Software Engineering}, year={2026}, doi={10.1007/s10664-026-10943-y} }

Empirical Software Engineering (EMSE), 2026
MetaMorphic Testing

Metamorphic Testing for Clinical ML Models: A Framework Proposal and Pilot Study

Jie JW Wu, Feiyu E, Bo Chen

Arxiv @article{wu2026metamorphic, title={Metamorphic Testing for Clinical ML Models: A Framework Proposal and Pilot Study}, author={Wu, Jie JW and E, Feiyu and Chen, Bo}, journal={arXiv preprint arXiv:2607.22984}, year={2026} }

3rd ACM International Conference on AI-powered Software (AIware 2026), Arxiv Track
GrowthHacker

GrowthHacker: Automated Off-Policy Evaluation Optimization Using Code-Modifying LLM Agents

Jie JW Wu*, Ayanda Patrick Herlihy*, Ahmad Saleem Mirza, Ali Afoud, Fatemeh Fard (*Equal Contribution)

Paper Code Blog @article{wuherlihy2025growthhacker, title={GrowthHacker: Automated Off-Policy Evaluation Optimization Using Code-Modifying LLM Agents}, author={Wu, Jie JW and Herlihy, Ayanda Patrick and Mirza, Ahmad Saleem and Afoud, Ali and Fard, Fatemeh}, journal={ACM Transactions on Software Engineering and Methodology}, year={2026} }

ACM Transactions on Software Engineering and Methodology (TOSEM), 2026

2025

FISCAL

FISCAL: Financial Synthetic Claim–document Augmented Learning for Efficient Fact-Checking

Rishab Sharma, Iman Saberi, Elham Alipour, Jie JW Wu, Fatemeh H. Fard

Arxiv Code @inproceedings{sharma2025fiscal, title={FISCAL: Financial Synthetic Claim–document Augmented Learning for Efficient Fact-Checking}, author={Sharma, Rishab and Saberi, Iman and Alipour, Elham and Wu, Jie JW and Fard, Fatemeh H.}, booktitle={NeurIPS 2025 Workshop on GenAI in Finance}, year={2025} }

NeurIPS 2025 Workshop "GenAI in Finance".
Technique deployed in Charli AI’s financial production system.
LLM Survey on Low Resource PL

A Survey on LLM-based Code Generation for Low-Resource and Domain-Specific Programming Languages

Sathvik Joel*, Jie JW Wu*, Fatemeh Hendijani Fard
(*Equal contribution)

GitHub @article{joel2024survey, title={A Survey on LLM-based Code Generation for Low-Resource and Domain-Specific Programming Languages}, author={Joel, Sathvik and Wu, Jie JW and Fard, Fatemeh H}, journal={arXiv preprint arXiv:2410.03981v3}, year={2024} }

ACM Transactions on Software Engineering and Methodology (TOSEM), 2025 (Presented at FSE 2026)
LLM Communication Benchmark

HumanEvalComm: Benchmarking the Communication Competence of Code Generation for LLMs and LLM Agent

Jie JW Wu, Fatemeh Hendijani Fard

Code Poster Blog Video @article{wu2025benchmarking, author = {Wu, Jie JW and Fard, Fatemeh H.}, title = {HumanEvalComm: Benchmarking the Communication Competence of Code Generation for LLMs and LLM Agent}, journal = {ACM Trans. Softw. Eng. Methodol.}, year = {2025}, doi = {10.1145/3715109}, url = {https://doi.org/10.1145/3715109} }

ACM Transactions on Software Engineering and Methodology (TOSEM), 2025 (Presented at FSE 2026)

2024

Autmoated Offline AB

AutoOffAB: Toward Automated Offline A/B Testing for Data-Driven Requirement Engineering

Jie JW Wu

32nd ACM International Conference on the Foundations of Software Engineering (FSE ’24, Ideas, Visions and Reflections Track)

Arxiv @inproceedings{wu2024autooffab, title={AutoOffAB: Toward Automated Offline A/B Testing for Data-Driven Requirement Engineering}, author={Wu, Jie JW}, booktitle={Companion Proceedings of the 32nd ACM International Conference on the Foundations of Software Engineering}, pages={472--476}, year={2024} }
Code Clone Detection

Investigating the Efficacy of Large Language Models for Code Clone Detection

Mohamad Khajezade, Jie JW Wu, Fatemeh Hendijani Fard, Gema Rodríguez-Pérez,Mohamed Sami Shehata

International Conference on Program Comprehension (ICPC 2024), Early Research Achievements (ERA) (short paper)

Arxiv @inproceedings{khajezade2024investigating, title={Investigating the Efficacy of Large Language Models for Code Clone Detection}, author={Khajezade, Mohamad and Wu, Jie JW and Fard, Fatemeh Hendijani and Rodr{\'\i}guez-P{\'e}rez, Gema and Shehata, Mohamed Sami}, booktitle={Proceedings of the 32nd IEEE/ACM International Conference on Program Comprehension}, pages={161--165}, year={2024} }
V-Model

An Exploratory Study of V-Model in Building ML-Enabled Software: A Systems Engineering Perspective

Jie JW Wu

3rd International Conference on AI Engineering – Software Engineering for AI (CAIN 2024) (long paper) Distinguished Paper Award Candidate.

Slides Blog @inproceedings{wu2024exploratory, title={An Exploratory Study of V-Model in Building ML-Enabled Software: A Systems Engineering Perspective}, author={Wu, Jie JW}, booktitle={3rd International Conference on AI Engineering – Software Engineering for AI (CAIN 2024)}, note={Lisbon, Portugal} }

2023

LLM Communication

Large Language Models Should Ask Clarifying Questions to Increase Confidence in Generated Code

Jie JW Wu

The 7th Annual Symposium on Machine Programming (MAPS ’23 Workshop)

Slides @inproceedings{wu2023large, title={Large Language Models Should Ask Clarifying Questions to Increase Confidence in Generated Code}, author={Wu, Jie JW}, booktitle={The 7th Annual Symposium on Machine Programming (MAPS ’23 Workshop)}, note={Workshop, December 3, 2023, San Francisco, CA, USA} }
MOVSW

A Multi-Objective Evolutionary Approach Towards Automated Online Controlled Experiments

Jie JW Wu, Thomas A. Mazzuchi, and Shahram Sarkani

Journal of Systems and Software, 2023.

Pre-print Code @article{wu2023multi, title={A multi-objective evolutionary approach towards automated online controlled experiments}, author={Wu, Jie JW and Mazzuchi, Thomas A and Sarkani, Shahram}, journal={Journal of Systems and Software}, pages={111703}, year={2023}, publisher={Elsevier} }
LDM

Comparison of Multi-Criteria Decision-Making Methods for Online Controlled Experiments in a Launch Decision-Making Framework

Jie JW Wu, Thomas A. Mazzuchi, and Shahram Sarkani

Information and Software Technology 155 (2023): 107115.

Pre-print Code Slides @article{wu2023comparison, title={Comparison of multi-criteria decision-making methods for online controlled experiments in a launch decision-making framework}, author={Wu, Jie JW and Mazzuchi, Thomas A and Sarkani, Shahram}, journal={Information and Software Technology}, volume={155}, pages={107115}, year={2023}, publisher={Elsevier} }

Earlier

Offline Sketch Parsing

Offline Sketch Parsing via Shapeness Estimation

Jie Wu, Changhu Wang, Liqing Zhang, Yong Rui

IJCAI. Vol. 15. 2015. (long paper, acceptance rate:28.6%)
Online Sketch Recognition

Sketch Recognition with Natural Correction and Editing

Jie Wu, Changhu Wang, Liqing Zhang, Yong Rui

Proceedings of the AAAI Conference on Artificial Intelligence. Vol. 28. No. 1. 2014. (long paper, acceptance rate:28.0%)

Video
Gestalt Saliency

Gestalt saliency: Salient region detection based on gestalt principles

Jie Wu, Liqing Zhang

2013 IEEE International Conference on Image Processing. IEEE, 2013. Top 1% Excellent Undergraduate Thesis Award at SJTU

← Back to Homepage