Publications

Under review.

Policy and World Modeling Co-Training for Language Agents

Ning Lu*, Baijiong Lin*, Shengcai Liu, Jiahao Wu, Haoze Lv, Yanbin Wei, Lingting Zhu, Shengju Qian, Xin Wang, Ying-Cong Chen, Qi Wang, Ke Tang (* equal contribution)

The first policy and world-modeling co-training RL framework for LLM agents.

Under review.

AHD Agent: Agentic Reinforcement Learning for Automatic Heuristic Design

Haoze Lv*, Ning Lu*, Ziang Zhou, Shengcai Liu (* equal contribution)

The first tool-integrated multi-turn agentic framework for automatic algorithm design.

Under review.

Train at the Moving Edge: Efficient RL for Large Reasoning Models via Rollout Selection

Jiahao Wu*, Ning Lu*, Shengcai Liu, Kun Wang, Yanting Yang, Li Qing, Ke Tang (* equal contribution)

The first online policy-verified data selection framework for efficient RL training.

IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)

VL-RouterBench: A Benchmark for Vision-Language Model Routing

Zhehao Huang, Baijiong Lin, Jingyuan Zhang, Jingying Wang, Yuhang Liu, Ning Lu, Tao Li, Xiaolin Huang

The first benchmark for VLM router.

Conference on Neural Information Processing Systems (NeurIPS)

Is PRM Necessary? Problem-Solving RL Implicitly Induces PRM Capability in LLMs

Zhangying Feng*, Qianglong Chen*, Ning Lu, Yongqian Li, Siqi Cheng, Shuangmu Peng, Duyu Tang, Shengcai Liu, Zhirui Zhang (* equal contribution)

Unifying problem solving and solution-process judgment.

International Conference on Machine Learning (ICML)

Safe Delta: Consistently Preserving Safety when Fine-Tuning LLMs on Diverse Datasets

Ning Lu, Shengcai Liu, Jiahao Wu, Weiyu Chen, Zhirui Zhang, Yew-Soon Ong, Qi Wang, Ke Tang

The first safety-aware post-fine-tuning defense method for LLM alignment.

Under review.

SemDiff: Generating Natural Unrestricted Adversarial Examples via Semantic Attributes Optimization in Diffusion Models

Zeyu Dai, Shengcai Liu, Rui He, Jiahao Wu, Ning Lu, Wenqi Fan, Qing Li, Ke Tang

IEEE International Conference on Data Engineering (ICDE) Oral

Backdoor graph condensation

Jiahao Wu, Ning Lu, Zeiyu Dai, Kun Wang, Wenqi Fan, Weiyu Chen, Shengcai Liu, Qing Li, Ke Tang

2024 IEEE 35th International Symposium on Software Reliability Engineering Workshops (ISSREW)

Training Overhead Ratio: A Practical Reliability Metric for Large Language Model Training Systems

Ning Lu, Qian Xie, Hao Zhang, Wenyi Fang, Yang Zheng, Zheng Hu, Jiantao Ma

2024 IEEE Conference on Artificial Intelligence (CAI) Oral

Less is More: Understanding Word-level Textual Adversarial Attack via n-gram Frequency Descend

Ning Lu, Shengcai Liu, Zhirui Zhang, Qi Wang, Haifeng Liu, Ke Tang

Transactions on Machine Learning Research (TMLR)

Large Language Models can be Guided to Evade AI-generated Text Detection

Ning Lu, Shengcai Liu, Rui He, Qi Wang, Yew-Soon Ong, Ke Tang

Showing that LLMs themselves can evade AI detectors with fine-grained prompting.

ACM Transactions on Evolutionary Learning and Optimization

Effective and Imperceptible Adversarial Textual Attack via Multi-objectivization

Shengcai Liu, Ning Lu, W Hong, C Qian, Ke Tang

IEEE/ACM Transactions on Audio, Speech, and Language Processing (TALSP, CCF-B)

Efficient Combinatorial Optimization for Word-level Adversarial Textual Attack

Shengcai Liu, Ning Lu, Cheng Chen, Ke Tang