Papers by Zhichao Xu
CrossGuard: Safeguarding MLLMs against Joint-Modal Implicit Malicious Attacks (2026.acl-long)
Copied to clipboard
| Challenge: | Existing methods for MLLMs are weak on explicit attacks, but weak on implicit ones. |
| Approach: | They propose an automated red-teaming pipeline that leverages reinforcement learning with tailored reward modules to generate diverse implicit samples across 14 domains. |
| Outcome: | The proposed method outperforms existing methods in implicit and explicit attacks while maintaining high utility. |
SLOT: Structuring the Output of Large Language Models (2025.emnlp-industry)
Copied to clipboard
| Challenge: | Structured outputs are essential for large language models (LLMs) but often deviate from predefined schemas hampering reliable application development. |
| Approach: | They propose a model-agnostic approach that transforms unstructured LLM outputs into precise structured formats. |
| Outcome: | The proposed model-agnostic approach transforms unstructured LLM outputs into precise structured formats. |
MAIN-RAG: Multi-Agent Filtering Retrieval-Augmented Generation (2025.acl-long)
Copied to clipboard
Chia-Yuan Chang, Zhimeng Jiang, Vineeth Rakesh, Menghai Pan, Chin-Chia Michael Yeh, Guanchu Wang, Mingzhi Hu, Zhichao Xu, Yan Zheng, Mahashweta Das, Na Zou
| Challenge: | Existing RAG systems struggle with the quality of retrieval documents, causing performance degradation and reducing performance. |
| Approach: | They propose a training-free RAG framework that leverages multiple LLM agents to collaboratively filter and score retrieved documents. |
| Outcome: | The proposed framework outperforms existing RAG frameworks in QA benchmarks and shows superior answer consistency and answer accuracy over baseline methods. |
A Systematic Survey of Automatic Prompt Optimization Techniques (2025.emnlp-main)
Copied to clipboard
Kiran Ramnath, Kang Zhou, Sheng Guan, Soumya Smruti Mishra, Xuan Qi, Zhengyuan Shen, Shuai Wang, Sangmin Woo, Sullam Jeoung, Yawei Wang, Haozhu Wang, Han Ding, Yuzhe Lu, Zhichao Xu, Yun Zhou, Balasubramaniam Srinivasan, Qiaojing Yan, Yueyan Chen, Haibo Ding, Panpan Xu, Lin Lee Cheong
| Challenge: | Recent advances in prompt engineering have created impediments for end users to adopt . however, prompt engineering remains an impedance due to rapid advances in models, tasks, and associated best practices. |
| Approach: | They propose to define APO as a 5-part unifying framework and categorize all relevant works based on their salient features. |
| Outcome: | The proposed framework aims to improve the performance of large language models on various tasks. |
In-Context Example Ordering Guided by Label Distributions (2024.findings-naacl)
Copied to clipboard
| Challenge: | Prior work has shown that ICL is sensitive to different natural language instructions and different orderings of in-context examples. |
| Approach: | They propose two principles for in-context example ordering guided by model’s probability predictions. |
| Outcome: | The proposed model outperforms baseline models on 13 text classification datasets and nine autoregressive LLMs with 700M to 13B parameters. |
IPR: Intelligent Prompt Routing with User-Controlled Quality-Cost Trade-offs (2025.emnlp-industry)
Copied to clipboard
Aosong Feng, Balasubramaniam Srinivasan, Yun Zhou, Zhichao Xu, Kang Zhou, Sheng Guan, Yueyan Chen, Xian Wu, Ninad Kulkarni, Yi Zhang, Zhengyuan Shen, Dmitriy Bespalov, Soumya Smruti Mishra, Yifei Teng, Darren Yow-Bang Wang, Haibo Ding, Lin Lee Cheong
| Challenge: | Existing systems require users to manually select models or employ rigid routing rules that fail to capture the continuous spectrum of query complexity. |
| Approach: | They propose a quality-constrained intelligent prompt routing framework that automatically selects optimal models based on predicted response quality and user-specified tolerance levels. |
| Outcome: | The proposed framework achieves 43.9% cost reduction while maintaining quality parity with strongest model in the Claude family and processes requests with sub-150ms latency. |
EDDA: An Encoder-Decoder Data Augmentation Framework for Zero-Shot Stance Detection (2024.lrec-main)
Copied to clipboard
| Challenge: | Existing methods for enhancing text or data are limited by lack of logical connections between generated texts and training data. |
| Approach: | They propose an encoder-decoder data augmentation framework that combines large language models and chain-of-thought prompting to summarize texts into target-specific if-then rationales, establishing logical relationships. |
| Outcome: | The proposed framework significantly improves over state-of-the-art methods on benchmark datasets while enabling interpretable rationale-based learning. |
Beyond Perplexity: Multi-dimensional Safety Evaluation of LLM Compression (2024.findings-emnlp)
Copied to clipboard
| Challenge: | Prior work on compression prioritizes preserving perplexity, which is analogous to training loss. |
| Approach: | They examine the impact of model compression along four dimensions: degeneration harm, representational harm, dialect bias, and language modeling and downstream task performance. |
| Outcome: | The proposed compression methods can lead to unexpected consequences, the authors show . quantization preserves bias while pruning degrades quickly. |
SALT: Step-level Advantage Assignment for Long-horizon Agents via Trajectory Graph (2026.findings-eacl)
Copied to clipboard
| Challenge: | Large Language Models (LLMs) have demonstrated remarkable capabilities, but their application to complex, multi-step, and long-horizon tasks remains challenging. |
| Approach: | They propose a framework that provides a finer-grained advantage assignment derived solely from outcome rewards. |
| Outcome: | The proposed framework provides a finer-grained advantage assignment, derived solely from outcome rewards. |
Diffusion Language Model Inference with Monte Carlo Tree Search (2026.findings-eacl)
Copied to clipboard
Zheng Huang, Kiran Ramnath, Yueyan Chen, Aosong Feng, Sangmin Woo, Balasubramaniam Srinivasan, Zhichao Xu, Kang Zhou, Shuai Wang, Haibo Ding, Lin Lee Cheong
| Challenge: | Existing methods for inference use heuristics to determine which positions to unmask and which tokens to commit . MEDAL is an inference-time scaling framework that integrates Monte Carlo Tree SEarch initialization for Diffusion Language Model inference. |
| Approach: | They propose a framework that integrates Monte Carlo Tree SEarch initialization for Diffusion Language Model inference. |
| Outcome: | The proposed framework achieves 22.0% improvement over existing inference strategies across multiple benchmarks. |
Reinforcement Learning for Self-Improving Agent with Skill Library (2026.acl-long)
Copied to clipboard
Jiongxiao Wang, Qiaojing Yan, Yawei Wang, Yijun Tian, Soumya Smruti Mishra, Zhichao Xu, Megha Gandhi, Panpan Xu, Lin Lee Cheong
| Challenge: | Large Language Model (LLM)-based agents have demonstrated remarkable capabilities in complex reasoning and multi-turn interactions but struggle to continuously improve and adapt when deployed in new environments. |
| Approach: | They propose a Reinforcement Learning-based approach to enhance agents’ self-improvement capabilities with a skill library. |
| Outcome: | The proposed framework achieves 8.9% higher Scenario Goal Completion when applied to supervised-finetuned model with expert experience while requiring 26% fewer interaction steps and generating 59% fewer tokens. |
Learning to Ideate for Machine Learning Engineering Agents (2026.eacl-short)
Copied to clipboard
Yunxiang Zhang, Kang Zhou, Zhichao Xu, Kiran Ramnath, Yun Zhou, Sangmin Woo, Haibo Ding, Lin Lee Cheong
| Challenge: | Existing machine learning engineering (MLE) agents struggle to iteratively optimize their implemented algorithms for effectiveness. |
| Approach: | They propose a framework that separates ideation from implementation that allows an implementation agent to request strategic help from a dedicated Ideator. |
| Outcome: | The proposed framework outperforms implementation-only agent baselines on MLE-Bench and can be trained with reinforcement learning to generate more effective ideas. |
NoteChat: A Dataset of Synthetic Patient-Physician Conversations Conditioned on Clinical Notes (2024.findings-acl)
Copied to clipboard
| Challenge: | NoteChat is a cooperative multi-agent framework for generating patient-physician dialogues . evaluator finds it outperforms state-of-the-art models for generating clinical notes . clinical documentation is largely done by physicians at both steps . |
| Approach: | They propose a cooperative multi-agent framework leveraging Large Language Models to generate patient-physician dialogues. |
| Outcome: | The proposed framework outperforms state-of-the-art models for generating clinical notes . it can engage patients directly and help clinical documentation, a leading cause of physician burnout . |
BayesFlow: A Probability Inference Framework for Meta-Agent Assisted Workflow Generation (2026.findings-eacl)
Copied to clipboard
| Challenge: | Large Language Models (LLMs) have demonstrated remarkable generality, often solving tasks with a single carefully engineered prompt. |
| Approach: | They propose to cast automatic workflow generation as Bayesian inference over a posterior distribution on workflows and instantiate BayesFlow as Bayer-based workflow generation framework. |
| Outcome: | The proposed framework improves accuracy by 9 percentage points over baselines and 65 percentage points on pool-wide benchmarks. |
FanChuan: A Multilingual and Graph-Structured Benchmark For Parody Detection and Analysis (2025.findings-acl)
Copied to clipboard
Yilun Zheng, Sha Li, Fangkun Wu, Yang Ziyi, Lin Hongchao, Zhichao Hu, Cai Xinjun, Ziming Wang, Jinxuan Chen, Sitao Luan, Jiahao Xu, Lihui Chen
| Challenge: | Parody is an emerging phenomenon on social media, where individuals imitate a role or position opposite to their own . limited available data and deficient diversity in current datasets hinder study of parody . |
| Approach: | They build a dataset of parody users and annotated comments from both English and Chinese corpora to test parody detection and comment sentiment analysis. |
| Outcome: | The proposed datasets provide richer contextual information, which is lacking in existing datasets. |
Unlocking the Power of Large Language Models for Entity Alignment (2024.acl-long)
Copied to clipboard
Xuhui Jiang, Yinghan Shen, Zhichao Shi, Chengjin Xu, Wei Li, Zixuan Li, Jian Guo, Huawei Shen, Yuanzhuo Wang
| Challenge: | Entity Alignment (EA) is a crucial step in unifying data from heterogeneous sources and plays a critical role in data-driven AI applications. |
| Approach: | They propose a framework that incorporates large language models to improve EA. |
| Outcome: | The proposed framework incorporates large language models (LLMs) to improve EA accuracy while preserving efficiency. |
Multi-dimensional Evaluation of Empathetic Dialogue Responses (2024.findings-emnlp)
Copied to clipboard
| Challenge: | Prior efforts to measure conversational empathy focus on expressed communicative intents, but ignore the fact that conversation is also a collaboration involving both speakers and listeners. |
| Approach: | They propose a multi-dimensional empathy evaluation framework to measure both expressed intents from the speaker’s perspective and perceived empathy from the listener’s viewpoint. |
| Outcome: | The proposed framework measures both expressed intents from the speaker’s perspective and perceived empathy from the listener’s viewpoint. |
SeqPO-SiMT: Sequential Policy Optimization for Simultaneous Machine Translation (2025.findings-acl)
Copied to clipboard
| Challenge: | SeqPO-SiMT is a new policy optimization framework for simultaneous machine translation that combines a tailored reward with a single step task. |
| Approach: | They propose a new policy optimization framework that defines the simultaneous machine translation task as a sequential decision making problem with a tailored reward. |
| Outcome: | The proposed framework outperforms the supervised fine-tuning model by 1.13 points while reducing the Average Lagging by 6.17 in the NEWSTEST2021 En Zh dataset. |
Sentiment Interpretable Logic Tensor Network for Aspect-Term Sentiment Analysis (2022.coling-1)
Copied to clipboard
| Challenge: | Aspect-term sentiment analysis (ATSA) is a fine-grained task that aims to infer the sentiment towards the given aspect-terms. |
| Approach: | They propose a novel ATSA method that is interpretable and has high accuracy . they propose SILTN, which is a neurosymbolic formalism, to improve the accuracy based on syntax knowledge distillation. |
| Outcome: | The proposed method is interpretable because it is a neurosymbolic formalism and a computational model that supports learning and reasoning about data with a differentiable first-order logic language. |
JudgeAgent: Beyond Static Benchmarks for Knowledge-Driven and Dynamic LLM Evaluation (2026.findings-acl)
Copied to clipboard
Zhichao Shi, Xuhui Jiang, Chengjin Xu, Cangli Yao, Shengjie Ma, Yinghan Shen, Zixuan Li, Jian Guo, Yuanzhuo Wang
| Challenge: | Current evaluation methods for large language models rely on static benchmarks . limited knowledge coverage and fixed difficulties hinder the targeted optimizations resulting in superficial evaluations of LLMs - a problem that has been addressed by JudgeAgent . |
| Approach: | They propose a knowledge-driven and dynamic evaluation framework for large language models . judgeAgent leverages LLM agents equipped with context graphs to traverse knowledge structures . |
| Outcome: | The proposed framework can achieve comprehensive evaluations and facilitate effective model iterations. |
MM-ChatAlign: A Novel Multimodal Reasoning Framework based on Large Language Models for Entity Alignment (2024.findings-emnlp)
Copied to clipboard
| Challenge: | Existing MMEA methods rely on knowledge representation learning (KRL) to measure the similarity of entity embeddings. |
| Approach: | They propose a framework that utilizes the visual reasoning abilities of MLLMs for multimodal entity alignment. |
| Outcome: | The proposed framework integrates the visual reasoning abilities of MLLMs for multimodal entity alignment. |
DataArc-SynData-Toolkit: A Unified Closed-Loop Framework for Multi-Path, Multimodal, and Multilingual Data Synthesis (2026.acl-demo)
Copied to clipboard
Zhichao Shi, Cehao Yang, Hao Zhou, Xiaojun Wu, Huajie Li, Xuhui Jiang, Chengjin Xu, Yuanzhuo Wang, Jian Guo
| Challenge: | Existing synthetic data tools are limited by convoluted workflows, fragmented data standards, and limited scalability across modalities. |
| Approach: | They develop an open-source framework that aims to reduce the technical barrier to synthetic data generation and subsequent model training. |
| Outcome: | The proposed framework achieves an optimal balance between generation efficiency and data quality. |