Papers by Weimin Lyu
Attention-Enhancing Backdoor Attacks Against BERT-based Models (2023.findings-emnlp)
Copied to clipboard
| Challenge: | Existing textual backdoor attacks focus on generating stealthy triggers or modifying model weights. |
| Approach: | They propose a Trojan Attention Loss (TAL) which enhances the Trojan behavior by directly manipulating attention patterns. |
| Outcome: | The proposed method improves the effectiveness of the backdoor attacks on different backbone models and tasks. |
OPeRA: A Dataset of Observation, Persona, Rationale, and Action for Evaluating LLMs on Human Online Shopping Behavior Simulation (2026.acl-long)
Copied to clipboard
Ziyi Wang, Yuxuan Lu, Wenbo Li, Amirali Amini, Bo Sun, Yakov Bart, Weimin Lyu, Jiri Gesi, Tian Wang, Jing Huang, Yu Su, Upol Ehsan, Malihe Alikhani, Toby Jia-Jun Li, Lydia Chilton, Dakuo Wang
| Challenge: | evaluating LLMs' ability to mimic real user behavior remains an open challenge due to the lack of high-quality, publicly available datasets that capture both the observable actions and the internal reasoning of an actual user. |
| Approach: | They propose a dataset of Observation, Persona, Rationale, and Action collected from real human participants during online shopping sessions. |
| Outcome: | The proposed dataset is the first to evaluate how well current LLMs can accurately simulate the next web action of a specific user. |
Task-Agnostic Detector for Insertion-Based Backdoor Attacks (2024.findings-naacl)
Copied to clipboard
| Challenge: | Existing methods for textual backdoor detection are task-specific and less effective beyond sentence classification. |
| Approach: | They propose a task-agnostic method for backdoor detection that leverages final layer logits and an efficient pooling technique. |
| Outcome: | TABDet can jointly learn from diverse task-specific models, demonstrating superior detection efficacy over traditional methods. |
Class Distillation with Mahalanobis Contrast: An Efficient Training Paradigm for Pragmatic Language Understanding Tasks (2025.acl-long)
Copied to clipboard
| Challenge: | Existing classifiers for detecting deviant language often come with significant computational cost and high data demands. |
| Approach: | They propose a class-disstillation paradigm that targets the core challenge: distilling a small, well-defined target class from a heterogeneous background. |
| Outcome: | The proposed training paradigm outperforms baselines and large language models on three benchmarks. |
A Study of the Attention Abnormality in Trojaned BERTs (2022.naacl-main)
Copied to clipboard
| Challenge: | In computer vision, the trigger can be a fixed pattern overlaid on the images or videos. |
| Approach: | They propose an attention-based Trojan detector to distinguish Trojaned models from clean ones by observing the attention focus drifting behavior of Trojanes. |
| Outcome: | The proposed detector is based on transformer’s attention and can distinguish Trojan models from clean ones. |
Towards a Design Guideline for RPA Evaluation: A Survey of Large Language Model-Based Role-Playing Agents (2025.findings-acl)
Copied to clipboard
| Challenge: | Role-Playing Agents (RPAs) are increasingly popular due to diverse task requirements and agent designs. |
| Approach: | They propose an evidence-based evaluation design guideline for LLM-based RPAs based on agent attributes, task attributes, and evaluation metrics. |
| Outcome: | The proposed evaluation design guideline is based on a systematic review of 1,676 papers published between Jan. 2021 and Dec. 2024. |