About me
Hi! My name is Aiwei Liu (刘瑷玮). I am a final-year Ph.D. student at the School of Software in Tsinghua University, where I am advised by Prof. Lijie Wen. Before that, I received my B.Eng. degree from the Software Institue in Nanjing University in 2020.
Currently, I’m serving as a Visiting Scholar at the UIC BDSC Lab, working under the supervision of Prof. Philip S. Yu (ACM Fellow).
Previously, I was a research intern at Apple’s AIML Group, where I worked under the supervision of Dr. Meng Cao.
Additionally, I was serving as a Visiting Scholar at CUHK MISC Lab, working under the supervision of Prof. Irwin King (IEEE Fellow).
🔬 Research
- Preventing and Detecting Misinformation Generated by Large Language Models (SIGIR 2024 Tutorial) [Home] [Paper][Conference Page] 1️⃣
Watermark for Large Language Models
- An Unforgeable Publicly Verifiable Watermark for Large Language Models (ICLR 2024) [Paper] [Code] 1️⃣
- A Semantic Invariant Robust Watermark for Large Language Models (ICLR 2024) [Paper] [Code]1️⃣
- A Survey of Text Watermarking in the Era of Large Language Models (ACM Computing Surveys) [Paper][机器之心] [Twitter] [Home] 1️⃣
- Can Watermarked LLMs be Identified by Users via Crafted Prompts? (Pre-print) [Paper] 1️⃣
- MarkLLM: An Open-Source Toolkit for LLM Watermarking (EMNLP 2024 Demo) [Paper] [机器之心] [Code]💡
- An Entropy-based Text Watermarking Detection Method (ACL 2024 Main) [Paper] [Code]💡
- Cross-lingual Consistency for Text Watermark (ACL 2024 Main) [Paper] [Code]💡
- WaterSeeker: Efficient Detection of Watermarked Segments in Large Documents (Pre-print) [Paper] 💡
Safety Alignment for Large Language Models
- Direct Large Language Model Alignment Through Self-Rewarding Contrastive Prompt Distillation (ACL 2024 Main) [Paper] [Apple Website]1️⃣
- TIS-DPO: Token-level Importance Sampling for Direct Preference Optimization With Estimated Weights (Pre-print) [Paper] 1️⃣
Adversarial Examples for Large Language Models
- Character-level White-Box Adversarial Attacks against Transformers via Attachable Subwords Substitution (EMNLP 2022 Main) [Paper] [Code]1️⃣
Semantic Parsing with Large Language Models
- Semantic Enhanced Text-to-SQL Parsing via Iteratively Learning Schema Linking Graph (SIGKDD 2022) [Paper] [Code]1️⃣
- Exploring the Compositional Generalization in Context Dependent Text-to-SQL Parsing (ACL 2023 Findings) [Paper] [Code]1️⃣
- A comprehensive evaluation of ChatGPT’s zero-shot Text-to-SQL capability (Pre-print) [Paper] [Code] 1️⃣
Fact Checking with Large Language Models
Retrieval-Augmented Large Language Models
- Entropy-Based Decoding for Retrieval-Augmented Large Language Models (MINT@NeurIPS2024)[Paper] 💡
- Refiner: Restructure Retrieval Content Efficiently to Advance Question-Answering Capabilities (EMNLP 2024 Findings)[Paper] 💡
1️⃣: Leading contribution (First Author) 💡: Insightful contribution
🔥 News
- 2024.10: 🎉🎉 Excited to announce the our paper: MarkLLM: An Open-Source Toolkit for LLM Watermarking is accepted by EMNLP 2024 Demo Track.
- 2024.09: 🎉 One paper about Retrieval-Augmented Large Language Models is accepted by EMNLP 2024.
- 2024.08: 🎉🎉 Excited to announce the our paper: “A Survey of Text Watermarking in the Era of Large Language Models” Paper is accepted by ACM Computing Surveys!
- 2024.08: Invited as a reviewer for ICLR 2025.
- 2024.08: 🎉🎉 Excited to announce the updated version of our paper: “A Survey of Text Watermarking in the Era of Large Language Models” Paper!
- 2024.05: 🎉🎉 One paper about Large Language Model Alignment is accepted by ACL 2024.
- 2024.05: 🎉🎉 Two papers about watermark for Large Language Models are accepted by ACL 2024.
- 2024.05: 🎉🎉 One paper about Document Relation Extraction is accepted by Findings of ACL 2024.
- 2024.04: 🎉🎉 Our tutorial proposal “Preventing and Detecting Misinformation Generated by Large Language Models” is accepted by SIGIR 2024. SIGIR 2024.
- 2024.04: Invited as a reviewer for ACMMM 2024.
- 2024.04: Invited as a reviewer for ACL ARR April.
- 2024.02: Invited as a reviewer for ACL ARR February.
- 2024.01: 🎉🎉 Two papers about watermark for Large Language Models are accepted by ICLR 2024.
📞 Contact
- 📧 Email:
- liuaw20@mails.tsinghua.edu.cn
- liuaiwei20@gmail.com
- 💬 Wechat:
- u839134412