Yun Lin
Tenured Track Associate Professor
Works on software engineering with LLMs, including requirements compilation for software generation, agent security and explainable AI.
Yun Lin's lab at SJTU works on large language models, reasoning & agents and trustworthy & safe ML. 72 papers since 2023, 20 at top venues, mostly USENIX Security and FSE. 12 PhD students and 9 master's students. 6 alumni, most often to Huawei. Applications close .
At a glance
- papers, 2023–2026
- 19/ year
- at T1 venues
- 28%
- PhD students
- 12
- MS students
- 9
- alumni listed
- 6
Papers by yearfrom the lab homepage and DBLP, 2023–2026
- 22 papers2026 · T1 6 · collecting
- 15 papers2025 · T1 6
- 16 papers2024 · T1 3
- 19 papers2023 · T1 5
T1 venuesOtherVenues: USENIX Security, FSE, ICLR, ICSE, KDD
Research focuspaper titles that use each topic's words
- Large language models15 papers
- software5 papers
- Reasoning & agents2 papers
- compilation1 paper
- Trustworthy & safe ML0 papers
Paper list
From the lab homepage and DBLP.
202622 papers
- —AlphaCu: A Transformation-Driven Synthesis Framework for LLM-Based GPU Kernel Generation (ASE 2026)
- —Amino acid-based Biological Age Clock and Its Implications for Human Health and Aging (Nature Communication 2026)
- —Compiling Large Multi-Modal Requirement Documents into Runnable Software Systems: From an Agentic Test-Driven Perspective (ISSTA 2026)
- ACL (T1)CuBridge: An LLM-Based Framework for Understanding and Reconstructing High-Performance Attention Kernels (opens in a new tab)
- —CuBridge: An LLM-Based Framework for Understanding and Reconstructing High-Performance Attention Kernels (ACL 2026)
- —DRIP: Defending Prompt Injection via Token-wise Representation Editing and Residual Fusion (CCS 2026)
- (OOPSLAEditFlow: Benchmarking and Optimizing Code Edit Recommendation Systems via Reconstruction of Developer Flows
- —Enhancing Agentic Code Localization with Traceability Recovered from Repository Evolution (NeurIPS 2026)
- AAAI (T1)Few-Shot Precise Event Spotting via Unified Multi-Entity Graph and Distillation
- —From Patches to Trajectories: Privileged Process Supervision for Software-Engineering Agents (NeurIPS 2026)
- FSE (T1)Generalizing Test Cases for Comprehensive Test Scenario Coverage
- —Generating Project-Specific Test Cases with Requirement Validation Intention (ISSTA 2026)
202515 papers
- ICLR (T1)$F^3Set$: Towards Analyzing Fast, Frequent, and Fine-grained Events from Videos
- (SCA Sample-Free Compilation Framework for Efficient Dynamic Tensor Computation
- SIGSOFT FSE CompanionAEGIS: An Agent-based Framework for Bug Reproduction from Issue Descriptions (opens in a new tab)
- USENIX Security (T1)Are CAPTCHAs Still Bot-hard? Generalized Visual CAPTCHA Solving with Agentic Vision Language Model (opens in a new tab)
- PACMPL (T1)Automated Program Refinement: Guide and Verify Code Large Language Model with Refinement Calculus (opens in a new tab)
- —Automated Program Refinement: Guide and Verify Code Large Language Model with Refinement Calculus (POPL 2025)
- FSE (T1)CAShift: Benchmarking Log-Based Cloud Attack Detection under Normality Shift
- J. Comput. Sci. Technol.CoEdPilot: Interactively Recommending Project-Wise Code Edits (opens in a new tab)
- ICML (T1)Defending LVLMs Against Vision Attacks through Partial-Perception Supervision
- —FSE 2025, industrial track
- USENIX Security (T1)Generalized Visual CAPTCHA Solving with Agentic Vision Language Model
- (TOSEMGUI Test Migration via Abstraction and Concretization
202416 papers
- Frontiers Comput. Sci.A prompt-based approach to adversarial example generation and robustness enhancement (opens in a new tab)
- WWW (T1)A Symbolic Rule Integration Framework with Logic Transformer for Inductive Relation Prediction (opens in a new tab)
- —A Symbolic Rule Integration Framework with Logic Transformer for Inductive Relation Prediction (WWW 2024)
- ISSTAC2D2: Extracting Critical Changes for Real-World Bugs with Dependency-Sensitive Delta Debugging (opens in a new tab)
- ISSTACoEdPilot: Recommending Code Edits with Learned Prior Edit Relevance, Project-wise Awareness, and Interactive Nature (opens in a new tab)
- —CoEdPilot: Recommending Code Edits with Learned Prior Edit Relevance, Project-wise Awareness, and Interactive Nature (ISSTA 2024)
- ASEDetecting and Explaining Anomalies Caused by Web Tamper Attacks via Building Consistency-based Normality (opens in a new tab)
- —Detecting and Explaining Anomalies Caused by Web Tamper Attacks via Building Consistency-based Normality (ASE 2024)
- —Extracting Critical Changes for Real-World Bugs with Dependency-Sensitive Delta Debugging (ISSTA 2024)
- Formal Aspects Comput.Introduction to the Special Collection from PRDC 2023 (opens in a new tab)
- USENIX Security (T1)Less Defined Knowledge and More True Alarms: Reference-based Phishing Detection without a Pre-defined Reference List (opens in a new tab)
- —Less Defined Knowledge and More True Alarms: Reference-based Phishing Detection without a Pre-defined Reference List (USENIX Security 2024)
202319 papers
- —API-Knowledge Aware Search-based Software Testing: Where, What and How (FSE 2023)
- ESEC/SIGSOFT FSEAPI-Knowledge Aware Search-Based Software Testing: Where, What, and How (opens in a new tab)
- KDD (T1)B2-Sampling: Fusing Balanced and Biased Sampling for Graph Contrastive Learning (opens in a new tab)
- —B2-Sampling: Fusing Balanced and Biased Sampling for Graph Contrastive Learning (KDD 2023)
- ASEBugMiner: Automating Precise Bug Dataset Construction by Code Evolution History Mining (opens in a new tab)
- ICSE (T1)DeepArc: Modularizing Neural Networks for the Model Maintenance (opens in a new tab)
- —DeepArc: Modularizing Neural Networks for the Model Maintenance (ICSE 2023)
- ESEC/SIGSOFT FSEDeepDebugger: An Interactive Time-Travelling Debugging Approach for Deep Classifiers (opens in a new tab)
- —DeepDebugger: An Interactive Time-Travelling Debugging Approach for Deep Classifiers (FSE 2023)
- IEEE Trans. Software Eng.Demystifying Performance Regressions in String Solvers (opens in a new tab)
- KDD (T1)Empower Post-hoc Graph Explanations with Information Bottleneck: A Pre-training and Fine-tuning Perspective (opens in a new tab)
- —Empower Post-hoc Graph Explanations with Information Bottleneck: A Pre-training and Fine-tuning Perspective (KDD 2023)
Where alumni went6 listed, 25% to industry
Industry 1Academia 3
Most often to
- Huawei
Also to
- NUS
- HKUST
Apply to SJTUFall 2027
Computer Science PhD (international applicants)
- Round
- Self-funded
- Opens
- —
- GRE
- See program page
- English
- See program page
- Fee
- See program page
Not yet confirmed for this intake.
Similar labsnearest by research summary
- USStanford
Christopher Potts
- Multilingual, translation & linguistics
- USStanford
Yejin Choi
- Large language models
- GBEdinburgh
Shay Cohen
- Multilingual, translation & linguistics
- GBUCL
Yao Lu
- Large language models
- GBEdinburgh
Alexandra Birch
- Multilingual, translation & linguistics
- USNYU
Kyunghyun Cho
- Large language models
Your notes
Save this lab to keep notes.
Data
Checked against the lab homepage.
- Collected
- Verified
- Papers
- The lab homepage and DBLP
- Members
- Lab homepage
- Alumni
- Lab homepage
Tiers are venue classifications, not lab rankings.
SJTUcloses