Papers by Ting Yang

Beware of Your Po! Measuring and Mitigating AI Safety Risks in Role-Play Fine-Tuning of LLMs (2025.acl-long)

Challenge:	Existing methods for identifying user geolocation suffer from a lack of interpretability on the corresponding results.
Approach:	They adopt influence functions to interpret the behavior of GNN-based models by identifying the importance of training users when predicting locations.
Outcome:	The proposed method provides meaningful explanations on prediction results and also uncovers the so-called "black-box" GNN-based models by investigating the effect of individual nodes.

Copied to clipboard

Weixiang Zhao, Yulin Hu, Yang Deng, Jiahe Guo, Xingyu Sui, Xinyang Han, An Zhang, Yanyan Zhao, Bing Qin, Tat-Seng Chua, Ting Liu

Consolidation or Adaptation? PRISM: Disentangling SFT and RL Data via Gradient Concentration (2026.acl-long)

Challenge:	Existing role-play fine-tuning techniques improve role adaptability but may degrade safety performance, especially for villainous characters.
Approach:	They propose safety-aware Role-Play Fine-Tuning (SaRFT) to balance role-playing capabilities and safety.
Outcome:	The proposed method outperforms state-of-the-art baselines under both LoRA and full-parameter fine-tuning settings.

Copied to clipboard

Yang Zhao, Yangou Ouyang, Xiao Ding, Hepeng Wang, Bibo Cai, Kai Xiong, Jinglong Gao, Zhouhao Sun, Li Du, Bing Qin, Ting Liu

Yang Zhao, Li Du, Xiao Ding, Yangou Ouyang, Hepeng Wang, Kai Xiong, Jinglong Gao, Zhouhao Sun, Dongliang Xu, Qing Yang, Dongchen Li, Bing Qin, Ting Liu

Challenge:	Existing data arbitration strategies for large language model training rely on surface-level heuristics that fail to diagnose intrinsic learning needs.
Approach:	They propose a framework that arbitrates data based on its degree of cognitive conflict with the model's existing knowledge.
Outcome:	Extensive experiments on WebShop and ALFWorld show that PRISM outperforms state-of-the-art hybrid methods while reducing computational costs by up to 3.22 .

Copied to clipboard

TACLR: A Scalable and Efficient Retrieval-based Method for Industrial Product Attribute Value Identification (2025.acl-long)

Challenge:	Existing methods for selecting training data from general datasets fail to account for the joint distribution of instructions, resulting in inefficient learning and suboptimal knowledge transfer.
Approach:	They propose a method that constructs a mixed gradient-based instruction graph to capture the joint distribution and interdependencies among instructions.
Outcome:	The proposed method outperforms existing methods on domain adaptation tasks and in complex, data-scarce scenarios.

Copied to clipboard

Yindu Su, Huike Zou, Lin Sun, Ting Zhang, Haiyang Yang, Chen Li Yu, David Lo, Qingheng Zhang, Shuguang Han, Jufeng Chen

LLMs Meet Isolation Kernel: Lightweight, Learning-free Binary Embeddings for Fast Retrieval (2026.findings-acl)

Challenge:	Existing methods for product attribute value identification face critical challenges . seller-provided attribute values are often incomplete or inaccurate .
Approach:	They propose a retrieval-based method that uses taxonomy-aware contrastive learning . they use product profiles and candidate values to encode and retrieve attributes based on similarity .
Outcome:	The proposed method is based on a taxonomy-aware, hard negative sampling and adaptive inference with dynamic thresholds.

Copied to clipboard

Zhibo Zhang, Yang Xu, Kai Ming Ting, Cam-Tu Nguyen

Semantic Text Processing

Challenge:	Large language models (LLMs) embeddings are typically high-dimensional, leading to substantial storage and retrieval overhead.
Approach:	They propose a learning-free method that transforms an LLM embedding into a binary embeddable using Isolation Kernel (IKE).
Outcome:	The proposed method performs 16.7 faster retrieval and 16 lower memory usage than the original LLM embeddings while maintaining comparable accuracy.

A Semi-Supervised Stable Variational Network for Promoting Replier-Consistency in Dialogue Generation (D19-1)

Copied to clipboard

Jinxin Chang, Ruifang He, Longbiao Wang, Xiangyu Zhao, Ting Yang, Ruifang Wang

Adversarial Training for Machine Reading Comprehension with Virtual Embeddings (2021.starsem-1)

Challenge:	Existing methods favor uninformative and non replier-specific responses due to lack of relevant information guidance.
Approach:	They propose to use a semi-supervised variable network to generate replier-specific responses . they use vMF as latent space to obtain stable KL performance .
Outcome:	The proposed model outperforms baseline models on two large conversation datasets and generates diverse and replier-specific responses.

Copied to clipboard

Ziqing Yang, Yiming Cui, Chenglei Si, Wanxiang Che, Ting Liu, Shijin Wang, Guoping Hu

Reasoning

Challenge:	Neural networks are vulnerable to adversarial examples that have been mixed with certain perturbations.
Approach:	They propose a novel adversarial training method that perturbs the embedding matrix instead of word vectors to differentiate the roles of passages and questions.
Outcome:	The proposed method is effective universally and further improves the performance of MRC tasks.

On Safety Risks in Experience-Driven Self-Evolving Agents (2026.findings-acl)

Copied to clipboard

Weixiang Zhao, Yichen Zhang, Yingshuo Wang, Yang Deng, Yanyan Zhao, Xuda Zhi, Yongbo Huang, Hao He, Wanxiang Che, Bing Qin, Ting Liu

Challenge:	Experience-driven self-evolution has emerged as a promising paradigm for improving the autonomy of large language model agents, yet its reliance on self-curated experience introduces underexplored safety risks.
Approach:	They investigate how experience accumulation and utilization in self-evolving agents affect safety performance across web-based and embodied environments.
Outcome:	The findings expose inherent limitations of current self-evolving agents and call for more principled strategies to ensure safe and reliable adaptation.

MPO: Multilingual Safety Alignment via Reward Gap Optimization (2025.acl-long)

Copied to clipboard

Weixiang Zhao, Yulin Hu, Yang Deng, Tongtong Wu, Wenxuan Zhang, Jiahe Guo, An Zhang, Yanyan Zhao, Bing Qin, Tat-Seng Chua, Ting Liu

Multilinguality

Challenge:	Existing preference learning methods for safety alignment are monolingual and struggle with noisy multilingual data.
Approach:	They propose a multilingual reward gaP optimization approach that leverages the well-aligned safety capabilities of the dominant language to improve safety alignment across multiple languages.
Outcome:	Extensive experiments on three LLMs, LLaMA-3.1, Gemma-2 and Qwen2.5, validate MPO’s efficacy in multilingual safety alignment without degrading general multilingual utility.

Best Student Forcing: A Simple Training Mechanism in Adversarial Language Generation (2020.lrec-1)

Copied to clipboard

Jonathan Sauder, Ting Hu, Xiaoyin Che, Goncalo Mordido, Haojin Yang, Christoph Meinel

Robustness in NLP

Challenge:	Language models trained with Maximum Likelihood Estimation (MLE) have been considered as a mainstream solution in Natural Language Generation (NLG) however, they are reportedly suffering from training instability and mode collapse, and therefore outperform conventional MLE models.
Approach:	They propose a method to improve Generative Adversarial Nets (GANs) using best student forcing and discriminators to increase training stability and sample diversity.
Outcome:	The proposed techniques outperform MLE models and outperformed existing approaches in terms of sample diversity and training stability.

SeCuRepair: Semantics-Aligned, Curriculum-Driven, and Reasoning-Enhanced Vulnerability Repair Framework (2026.acl-long)

Copied to clipboard

Chengran Yang, Ting Zhang, Jinfeng Jiang, Xin Zhou, Haoye Tian, Mingzhe Du, Jieke Shi, Junkai Chen, Yikun Li, Eng Lieh Ouh, Lwin Khin Shar, David Lo

Reasoning

Challenge:	Existing methods for automating vulnerability repair suffer from syntactic overfitting . nvd published 49,230 Common Vulnerabilities and Exposures (CVE) records in 2025 alone .
Approach:	They propose a semantic-aware reward framework that optimizes for code semantic equivalence rather than lexical mimicry.
Outcome:	The proposed framework outperforms state-of-the-art frameworks on repository-level splits . it incorporates expert-aligned reasoning mechanism that grounds patch generation in structured diagnosis.

MAESTRO: Meta-learning Adaptive Estimation of Scalarization Trade-offs for Reward Optimization (2026.acl-long)

Copied to clipboard

Yang Zhao, Hepeng Wang, Xiao Ding, Yangou Ouyang, Bibo Cai, Kai Xiong, Jinglong Gao, Zhouhao Sun, Li Du, Bing Qin, Ting Liu

Adapting Where It Matters: Depth-Aware Adaptation for Efficient Multilingual Speech Recognition in Low-Resource Languages (2026.findings-acl)

Challenge:	Group-Relative Policy Optimization (GRPO) has emerged as an efficient paradigm for aligning Large Language Models (LLMs), but its efficacy is confined to domains with verifiable ground truths.
Approach:	They propose a meta-cognitive orchestration layer that treats reward scalarization as a dynamic latent policy, leveraging the model’s terminal hidden states as 'a semantic bottleneck' . Across seven benchmarks, MAESTRO consistently outperforms single-reward and static multi-objective baselines while preserving the efficiency advantages of GRPO.
Outcome:	The proposed model outperforms single-reward and static multi-objective baselines while preserving efficiency advantages.

Copied to clipboard

Yang Xiao, Eun-Jung Holden, Ting Dang

A Sentence Cloze Dataset for Chinese Machine Reading Comprehension (2020.coling-main)

Challenge:	Recent speech foundation models excel at multilingual automatic speech recognition (ASR) for high-resource languages, but their performance drops substantially on low-resourced languages due to the limited data availability.
Approach:	They propose a Depth-Aware Model Adaptation framework that allocates adaptation capacity according to each layer’s role.
Outcome:	The proposed framework matches or surpasses state-of-the-art accuracy with 80% fewer trainable parameters and achieves 29% error reduction under extreme data scarcity.

Copied to clipboard

Yiming Cui, Ting Liu, Ziqing Yang, Zhipeng Chen, Wentao Ma, Wanxiang Che, Shijin Wang, Guoping Hu

Reasoning

Challenge:	Using cloze-style reading comprehension, Chinese machine reading comprehension datasets are becoming more and more popular . a new task is proposed to fill the right candidate sentence into the passage with several blanks .
Approach:	They propose a Chinese task to fill the right candidate sentence into a passage with blanks . they build a dataset to evaluate the difficulty of the task and make fake candidates .
Outcome:	The proposed task fills the right candidate sentence into the passage with blanks . the proposed dataset contains over 100K blanks within over 10K passages based on Chinese narrative stories .

TextBrewer: An Open-Source Knowledge Distillation Toolkit for Natural Language Processing (2020.acl-demos)

Copied to clipboard

Ziqing Yang, Yiming Cui, Zhipeng Chen, Wanxiang Che, Ting Liu, Shijin Wang, Guoping Hu

One for All: Update Parameterized Knowledge Across Multiple Models with Once Edit (2025.acl-long)

Challenge:	Large pre-trained language models have hundreds of millions of parameters and take several gigabytes of memory to train and inference.
Approach:	They propose an open-source knowledge distillation toolkit designed for natural language processing that provides a set of predefined distillation methods and can be extended with custom code.
Outcome:	The proposed method is comparable with or even higher than the public distilled BERT models with similar numbers of parameters.

Copied to clipboard

Weitao Ma, Xiyuan Du, Xiaocheng Feng, Lei Huang, Yichong Huang, Huiyi Zhang, Xiaoliang Yang, Baohang Li, Xiachong Feng, Ting Liu, Bing Qin

Measuring Your ASTE Models in The Wild: A Diversified Multi-domain Dataset For Aspect Sentiment Triplet Extraction (2023.findings-acl)

Challenge:	Existing methods for modifying large language models focus on individual models, resulting in errors and hallucinations.
Approach:	They propose an ensemble-based approach that employs a plug-in model as the editing module and a dynamic weight mechanism to enhance its effectiveness.
Outcome:	The proposed approach outperforms existing methods while achieving superior editing efficiency.

Copied to clipboard

Ting Xu, Huiyun Yang, Zhen Wu, Jiaze Chen, Fei Zhao, Xinyu Dai

Sentiment Analysis

Challenge:	Existing ASTE datasets are limited in their ability to represent real-world scenarios, hindering progress in this area.
Approach:	They propose a new ASTE dataset that is manually annotated to better fit real-world scenarios by providing more diverse and realistic reviews.
Outcome:	The proposed dataset is manually annotated to better fit real-world scenarios.

Knowledge-Augmented Question Error Correction for Chinese Question Answer System with QuestionRAG (2025.emnlp-industry)

Copied to clipboard

Longpeng Qiu, Ting Li, Shuai Mao, Nan Yang, Xiaohui Yan

Explainability & Interpretability in NLP

Challenge:	Large language models struggle with input errors, often failing to interpret user intent or altering the original question’s structure (over-correction).
Approach:	They propose a framework that uses reinforcement learning to address misinterpretation and over-correction by integrating external knowledge with the input.
Outcome:	The proposed framework unlocks the full potential of LLMs for the question correction task.

Causal-Debias: Unifying Debiasing in Pretrained Language Models and Fine-tuning via Causal Invariant Learning (2023.acl-long)

Copied to clipboard

Fan Zhou, Yuzhou Mao, Liu Yu, Yi Yang, Ting Zhong

Robustness in NLP

Challenge:	Existing methods to remove unwanted stereotypical associations from pretrained language models (PLMs) are often focused on removing unwanted stereotypes from PLMs.
Approach:	They propose a framework to remove unwanted stereotypical associations in pretrained language models . they propose bias-relevant factors are causal, while labelrelevant factors causal .
Outcome:	The proposed framework reduces stereotypical associations after PLMs are fine-tuned . the proposed framework mitigates bias from a causal invariant perspective .

Chain of Strategy Optimization Makes Large Language Models Better Emotional Supporter (2025.findings-emnlp)

Copied to clipboard

Weixiang Zhao, Xingyu Sui, Xinyang Han, Yang Deng, Yulin Hu, Jiahe Guo, Libo Qin, Qianyun Du, Shijin Wang, Yanyan Zhao, Bing Qin, Ting Liu

Chase: A Large-Scale and Pragmatic Chinese Dataset for Cross-Database Context-Dependent Text-to-SQL (2021.acl-long)

Challenge:	Existing supervised fine-tuning (SFT) fails to address these issues, as it trains models on single gold-standard responses without modeling nuanced strategy trade-offs.
Approach:	They propose a two-stage framework that optimizes strategy selection preferences at each dialogue turn.
Outcome:	The proposed framework improves strategy selection preferences at each dialogue turn.

Copied to clipboard

Jiaqi Guo, Ziliang Si, Yu Wang, Qian Liu, Ming Fan, Jian-Guang Lou, Zijiang Yang, Ting Liu

Programming Languages in NLP

Challenge:	XDTS is a cross-database context-dependent text-to-sql problem with wide range of applications.
Approach:	They present a large-scale Chinese dataset for cross-database context-dependent Text-to-SQL . they find that only 35% of questions are context-independent and 28% of SQL queries are easy .
Outcome:	The proposed approach achieves an exact match accuracy of 40% over all questions and 16% over all question sequences.

Length Extrapolation of Transformers: A Survey from the Perspective of Positional Encoding (2024.findings-emnlp)

Copied to clipboard

Liang Zhao, Xiachong Feng, Xiaocheng Feng, Weihong Zhong, Dongliang Xu, Qing Yang, Hongtao Liu, Bing Qin, Ting Liu

AdaSteer: Your Aligned LLM is Inherently an Adaptive Jailbreak Defender (2025.emnlp-main)

Challenge:	Existing methods to enhance length extrapolation of large language models have been developed, but a systematic survey is lacking.
Approach:	They propose to examine the effects of positional encoding on length extrapolation.
Outcome:	The proposed methods improve the extrapolation of large language models, but they are still lacking a systematic survey.

Copied to clipboard

Weixiang Zhao, Jiahe Guo, Yulin Hu, Yang Deng, An Zhang, Xingyu Sui, Xinyang Han, Yanyan Zhao, Bing Qin, Tat-Seng Chua, Ting Liu

MURAL: Multimodal, Multitask Representations Across Languages (2021.findings-emnlp)

Challenge:	Activation steering offers training-free defense but relies on fixed steering coefficients, resulting in suboptimal protection and increased false rejections of benign inputs.
Approach:	They propose an adaptive activation steering method that dynamically adjusts model behavior based on input characteristics.
Outcome:	The proposed method outperforms baseline methods across multiple jailbreak attacks with minimal impact on utility.

Copied to clipboard

Aashi Jain, Mandy Guo, Krishna Srinivasan, Ting Chen, Sneha Kudugunta, Chao Jia, Yinfei Yang, Jason Baldridge

Semantic Text Processing

Challenge:	Image-caption pairs and translation pairs provide the means to learn deep representations of and connections between languages.
Approach:	They propose a dual encoder that integrates image-text matching and translation pairs to solve two tasks by learning from billions of pairs.
Outcome:	The proposed encoder outperforms ALIGN's cross-modal retrieval performance on well-resourced languages and significantly improves on under-resource languages.

Deciphering the Impact of Pretraining Data on Large Language Models through Machine Unlearning (2024.findings-acl)

Copied to clipboard

Yang Zhao, Li Du, Xiao Ding, Kai Xiong, Zhouhao Sun, Shi Jun, Ting Liu, Bing Qin

Analyzing the Rapid Generalization of SFT via the Perspective of Attention Head Activation Patterns (2025.acl-long)

Challenge:	Existing studies have suggested that the composition of the pretraining corpus exerts a significant impact upon the performance of LLMs.
Approach:	They analyze the impact of 48 datasets from 5 major categories of pretraining data of Large Language Models and measure their impacts on LLMs using benchmarks about nine major categories.
Outcome:	The proposed analysis provides insights into the organization of data to support more efficient pretraining of Large Language Models.

Copied to clipboard

Yang Zhao, Li Du, Xiao Ding, Kai Xiong, Ting Liu, Bing Qin

Benchmarking Robustness of Machine Reading Comprehension Models (2021.findings-acl)

Challenge:	Currently, LLMs learn in a data-driven schema while the instructions about complex tasks are both scarce and hard to collect or construct.
Approach:	They employ a gradient-based method to dissect the process that the Supervised Fine-tuning Process (SFT) adapts LLMs to downstream tasks via the perspective of attention patterns.
Outcome:	The proposed method dissects the process that the SFT process adapts LLMs to downstream tasks via the perspective of attention patterns.

Copied to clipboard

Chenglei Si, Ziqing Yang, Yiming Cui, Wentao Ma, Ting Liu, Shijin Wang

Robustness in NLP

Challenge:	Existing benchmarks only evaluate models' robustness under test-time perturbations or adversarial attacks.
Approach:	They propose a model-agnostic benchmark to evaluate models' robustness under adversarial attacks.
Outcome:	The proposed model-agnostic benchmark evaluates models under four different types of adversarial attacks.

SemRegex: A Semantics-Based Approach for Generating Regular Expressions from Natural Language Specifications (D18-1)

Copied to clipboard

Zexuan Zhong, Jiaqi Guo, Wei Yang, Jian Peng, Tao Xie, Jian-Guang Lou, Ting Liu, Dongmei Zhang

Multimodality

Challenge:	Existing approaches to generate programs from natural language do not address program aliasing . semantically equivalent programs may have many syntactically different forms .
Approach:	They propose a semantics-based approach to generate regular expressions from natural language.
Outcome:	The proposed approach improves on three public datasets.

Large Language Models Are Still Misled by Simple Bias Ensembles (2026.findings-acl)

Copied to clipboard

Zhouhao Sun, Zhiyuan Kan, Xiao Ding, Li Du, Bibo Cai, Yang Zhao, Bing Qin, Ting Liu

Online Iterative Self-Alignment for Radiology Report Generation (2025.acl-long)

Challenge:	Existing benchmarks for large language models are constrained to datasets where each sample is manually injected with only one type of bias.
Approach:	They propose a multi-bias benchmark where each sample contains multiple types of biases.
Outcome:	The proposed benchmark shows that existing LLMs and debiasing methods perform poorly on this benchmark, highlighting the challenge of eliminating compounded biases.

Copied to clipboard

Ting Xiao, Lei Shi, Yang Zhang, HaoFeng Yang, Zhe Wang, Chenjia Bai

Text Generation

Challenge:	Existing methods for RRG rely on supervised fine-tuning based on data pairs of radiological images and corresponding radiologist-annotated reports.
Approach:	They propose a method that performs supervised fine-tuning on data pairs of radiological images and corresponding radiologist-annotated reports.
Outcome:	The proposed method surpasses existing methods and achieves state-of-the-art performance across multiple evaluation metrics.

Knowledge Transfer in Incremental Learning for Multilingual Neural Machine Translation (2023.acl-long)

Copied to clipboard

Kaiyu Huang, Peng Li, Jin Ma, Ting Yao, Yang Liu

Multilinguality

Challenge:	Existing studies focus on overcoming catastrophic forgetting on original language pairs while lacking encouragement to learn new knowledge from incremental learning.
Approach:	They propose a knowledge transfer method that can adapt original MNMT models to diverse incremental language pairs by flexibly introducing knowledge from external models into original models, which encourages the models to learn new language pairs.
Outcome:	The proposed method outperforms baselines on multiple languages while maintaining performance on original language pairs.

SecureVibeBench: Benchmarking Secure Vibe Coding of AI Agents via Reconstructing Vulnerability-Introducing Scenarios (2026.acl-long)

Copied to clipboard

Junkai Chen, Huihui Huang, Yunbo Lyu, Junwen An, Jieke Shi, Chengran Yang, Ting Zhang, Haoye Tian, Yikun Li, Zhenhao Li, Xin Zhou, Xing Hu, David Lo

Challenge:	Existing benchmarks fail to capture scenarios in which vulnerabilities are introduced by humans . we evaluate 5 popular code agents supported by 5 LLMs on SecureVibeBench .
Approach:	They propose a benchmarking tool that compares 105 C/C++ secure coding tasks . they use real-world open-source vulnerabilities and a comprehensive evaluation tool .
Outcome:	The proposed benchmarks show that code agents struggle to produce correct and secure code . the best performing agent produces merely 23.8% correct and secured solutions .

DisCo: Distilled Student Models Co-training for Semi-supervised Text Mining (2023.emnlp-main)

Copied to clipboard

Weifeng Jiang, Qianren Mao, Chenghua Lin, Jianxin Li, Ting Deng, Weiyi Yang, Zheng Wang