Papers by Xinyi Wang

Gödel Agent: A Self-Referential Agent Framework for Recursively Self-Improvement (2025.acl-long)

Challenge:	Early debugging efforts focused on code-level analysis, which often fails when addressing complex programming errors.
Approach:	They propose a framework that employs natural language as an intermediate representation to improve code debugging by debuggating at a natural language level.
Outcome:	The proposed framework outperforms traditional debugging methods and enables a broader modification space through direct refinement guided by execution feedback.

Copied to clipboard

Xunjian Yin, Xinyi Wang, Liangming Pan, Li Lin, Xiaojun Wan, William Yang Wang

PAI-Diffusion: Constructing and Serving a Family of Open Chinese Diffusion Models for Text-to-image Synthesis on the Cloud (2024.acl-demos)

Challenge:	Existing agentic systems cannot search the whole design space due to the restriction of human-designed components.
Approach:	They propose a Gödel Agent framework that allows agents to recursively improve themselves without relying on fixed algorithms or fixed algorithms.
Outcome:	The proposed framework surpasses manual crafted agents in performance, efficiency, and generalizability.

Copied to clipboard

Chengyu Wang, Zhongjie Duan, Bingyan Liu, Xinyi Zou, Cen Chen, Kui Jia, Jun Huang

Disentangling Memory and Reasoning Ability in Large Language Models (2025.acl-long)

Challenge:	Existing diffusion models fail to address the challenges of generating high-quality images from textual descriptions due to its large vocabulary size and complex character relationships.
Approach:	They propose a framework that integrates Chinese diffusion models with Alibaba Cloud's Platform for AI and enables the generation of contextually relevant images.
Outcome:	The proposed framework integrates with Alibaba Cloud’s Platform for AI, providing accessible and scalable solutions.

Copied to clipboard

Mingyu Jin, Weidi Luo, Sitao Cheng, Xinyi Wang, Wenyue Hua, Ruixiang Tang, William Yang Wang, Yongfeng Zhang

ChatMap: Mining Human Thought Processes for Customer Service Chatbots via Multi-Agent Collaboration (2025.findings-acl)

Challenge:	Existing LLMs operate as an opaque process without explicit separation between knowledge retrieval and reasoning steps, making the decision-making process unclear and disorganized.
Approach:	They propose a language model inference paradigm that decomposes the complex inference process into two distinct and clear actions: (1) memory recall: which retrieves relevant knowledge, and (2) reasoning: which performs reasoning steps based on the recalled knowledge.
Outcome:	The proposed paradigm decomposes the inference process into two distinct and clear actions, memory and reason, guiding the model to distinguish between steps that require knowledge retrieval and those that involve reasoning.

Copied to clipboard

Xinyi Jiang, Tianyi Hu, Yuheng Qin, Guoming Wang, Zhou Huan, Kehan Chen, Gang Huang, Rongxing Lu, Siliang Tang

Dialogue Systems & Conversational Agents

Challenge:	Existing methods for enhancing dialogue performance rely on summarizing behavior . e-commerce chatbots need to align their dialogue strategies with human behavior to achieve coherent, human-like conversations with customers.
Approach:	They propose a method to extract core patterns from dialogue data and integrate them into models by mining service thought processes using a multi-agent aPproach.
Outcome:	The proposed method outperforms manual methods and outperfies baselines on Taobao in China.

Label Representations in Modeling Classification as Text Generation (2020.aacl-srw)

Copied to clipboard

Xinyi Chen, Jingxian Xu, Alex Wang

Information Retrieval

Challenge:	Existing methods for text generation use strings to represent labels . linguistic properties of labels do affect performance, though their results are limited to document retrieval.
Approach:	They investigate the effect of string representations on how effectively a model learns a task . they use four standard text classification tasks to model string representation .
Outcome:	The proposed model improves on four standard text classification tasks . the results are largely negative in the low data setting .

Uncovering Scaling Laws for Large Language Models via Inverse Problems (2025.findings-emnlp)

Copied to clipboard

Arun Verma, Zhaoxuan Wu, Zijian Zhou, Xiaoqiang Lin, Zhiliang Chen, Rachael Hwee Ling Sim, Rui Qiao, Jingtan Wang, Nhung Bui, Xinyuan Niu, Wenyang Hu, Gregory Kang Ruey Lau, Zi-Yu Khoo, Zitong Zhao, Xinyi Xu, Apivich Hemachandra, See-Kiong Ng, Bryan Kian Hsiang Low

A Tree-based Decoder for Neural Machine Translation (D18-1)

Challenge:	Large Language Models (LLMs) have achieved remarkable success across diverse domains.
Approach:	inverse problems can efficiently uncover scaling laws that guide the building of LLMs, authors argue . authors propose brute-force approaches to improve LLM training costs due to high costs .
Outcome:	This paper advocates that inverse problems can efficiently uncover scaling laws that guide the building of LLMs to achieve the desirable performance with significantly better cost-effectiveness.

Copied to clipboard

Xinyi Wang, Hieu Pham, Pengcheng Yin, Graham Neubig

DebateCoder: Towards Collective Intelligence of LLMs via Test Case Driven LLM Debate for Code Generation (2025.acl-long)

Challenge:	Existing work on adding syntactic information to NMT systems is limited to linguistically-inspired tree structures.
Approach:	They propose an NMT model that can naturally generate the topology of an arbitrary tree structure on the target side.
Outcome:	The proposed model outperforms standard seq2seq models by 2.1 BLEU points and other methods for incorporating target-side syntax by 0.7 BLUE points.

Copied to clipboard

Jizheng Chen, Kounianhua Du, Xinyi Dai, Weiming Zhang, Xihuai Wang, Yasheng Wang, Ruiming Tang, Weinan Zhang, Yong Yu

The Role of Model Confidence on Bias Effects in Measured Uncertainties for Vision-Language Models (2025.findings-emnlp)

Challenge:	Existing debate-based approaches to code generation are limited due to several reasons: 1) Reliance on different instances of the same LLM for debate, 2) under-utilization of test cases, and 3) reliance on third-party moderators for result consolidation and decision-making.
Approach:	They propose to use test cases to analyze code and identify bugs while opposing models generate test cases for each other to challenge each other's code during the debate process.
Outcome:	The proposed model collects intelligence of LLMs via test case-driven debate for code generation.

Copied to clipboard

Xinyi Liu, Weiguang Wang, Hangfeng He

Towards Rationality in Language and Multimodal Agents: A Survey (2025.naacl-long)

Challenge:	Quantifying epistemic uncertainty in open-ended tasks is challenging due to the presence of aleatoric uncertainty, which arises from multiple valid answers.
Approach:	They conduct experiments on visual question answering tasks and find that mitigating prompt-introduced bias improves uncertainty quantification.
Outcome:	The proposed approach reduces uncertainty quantification in visual question answering tasks by mitigating prompt-introduced biases.

Copied to clipboard

Bowen Jiang, Yangxinyu Xie, Xiaomeng Wang, Yuan Yuan, Zhuoqun Hao, Xinyi Bai, Weijie J Su, Camillo Jose Taylor, Tanwi Mallick

LLMTreeRec: Unleashing the Power of Large Language Models for Cold-Start Recommendations (2025.coling-main)

Challenge:	despite advances in language and multimodal agents, large language models lack rationality . despite their progress, large-scale models lack real-world grounding and feedback mechanisms .
Approach:	They propose to build more rational language and multimodal agents . they also examine what criteria define rationality in intelligent systems .
Outcome:	This paper assesses the state-of-the-art in language and multimodal agents . it also outlines open challenges and future research directions .

Copied to clipboard

Wenlin Zhang, Chuhan Wu, Xiangyang Li, Yuhao Wang, Kuicai Dong, Yichao Wang, Xinyi Dai, Xiangyu Zhao, Huifeng Guo, Ruiming Tang

Evaluating and Modeling Attribution for Cross-Lingual Question Answering (2023.emnlp-main)

Challenge:	Lack of training data leads to the system cold-start problem in recommendation systems, making them struggle to provide effective recommendations.
Approach:	They propose a tree-based LLM recommendation framework which structures all items into an item tree to improve the efficiency of LLM’s item retrieval.
Outcome:	The proposed framework outperforms the baseline model in the A/B test on Huawei industrial system.

Copied to clipboard

Benjamin Muller, John Wieting, Jonathan Clark, Tom Kwiatkowski, Sebastian Ruder, Livio Soares, Roee Aharoni, Jonathan Herzig, Xinyi Wang

Natural Language Interfaces

Challenge:	Open-retrieval question answering systems are lacking in attribution for cross-lingual question answering . open-research questions are available in 20 languages, but their raw generation often falls short in factuality .
Approach:	They are the first to study attribution for cross-lingual question answering . they collect data in 5 languages to assess the attribution level of a state-of-the-art QA system .
Outcome:	The proposed approach improves the attribution level of a state-of-the-art cross-lingual QA system.

Balancing Training for Multilingual Neural Machine Translation (2020.acl-main)

Copied to clipboard

Xinyi Wang, Yulia Tsvetkov, Graham Neubig

Humanity’s Last Code Exam: Can Advanced LLMs Conquer Human’s Hardest Code Competition? (2025.findings-emnlp)

Challenge:	Existing methods to train multilingual machine translation models are imbalanced and heterogeneous data is wildly varying.
Approach:	They propose a method that automatically learns how to weight training data through a data scorer that is optimized to maximize performance on all test languages.
Outcome:	The proposed method outperforms baselines on two sets of languages under one-to-many and many-to-1 MT settings and offers flexible control over which languages are optimized.

Copied to clipboard

Xiangyang Li, Xiaopeng Li, Kuicai Dong, null Zhangquanhu, Rongju Ruan, Xinyi Dai, Yasheng Wang, Ruiming Tang

TheoremQA: A Theorem-driven Question Answering Dataset (2023.emnlp-main)

Challenge:	o4-mini(high) and Gemini-2.5 Pro achieve pass@1 rates of only 15.9% and 11.4%, respectively.
Approach:	They propose a harmonized online–offline sandbox that guarantees fully reproducible evaluation.
Outcome:	The proposed test reflects the advanced reasoning and code generation ability of large language models.

Copied to clipboard

Wenhu Chen, Ming Yin, Max Ku, Pan Lu, Yixin Wan, Xueguang Ma, Jianyu Xu, Xinyi Wang, Tony Xia

2INER: Instructive and In-Context Learning on Few-Shot Named Entity Recognition (2023.findings-emnlp)

Challenge:	Recent LLMs like GPT-4 and PaLM-2 have made tremendous progress in solving fundamental math problems like GSM8K by achieving over 90% accuracy.
Approach:	They propose to use theorem-driven question-answering dataset to evaluate AI models' ability to apply theoretic concepts to solving challenging science problems.
Outcome:	TheoremQA is curated by domain experts and contains 800 high-quality questions covering 350 theoremics from Math, Physics, EE&CS, and Finance.

Copied to clipboard

Jiasheng Zhang, Xikai Liu, Xinyi Lai, Yan Gao, Shusen Wang, Yao Hu, Yiqing Lin

Low-Resource NLP

Challenge:	Named Entity Recognition (NER) tasks are a fundamental task of natural language processing (NLP).
Approach:	They propose a text-to-text framework for Few-Shot Named Entity Recognition (NER) that employs instruction finetuning and auxiliary tasks to enhance the model's understanding of entity types in the overall semantic context of a sentence.
Outcome:	The proposed framework outperforms existing Few-Shot NER methods and remains competitive with state-of-the-art NER algorithms.

SwitchOut: an Efficient Data Augmentation Algorithm for Neural Machine Translation (D18-1)

Copied to clipboard

Xinyi Wang, Hieu Pham, Zihang Dai, Graham Neubig

Multi-view Subword Regularization (2021.naacl-main)

Challenge:	Existing methods for data augmentation for text-based tasks such as machine translation are limited due to noise and noise.
Approach:	They propose a data augmentation policy with desirable properties as an optimization problem and propose 'SwitchOut' switchout randomly replaces words in both the source and target sentences with other random words from their corresponding vocabularies.
Outcome:	The proposed method outperforms strong alternatives such as word dropout on three translation datasets.

Copied to clipboard

Xinyi Wang, Sebastian Ruder, Graham Neubig

Efficient Test Time Adapter Ensembling for Low-resource Language Varieties (2021.findings-emnlp)

Challenge:	Existing subword regularization methods for multilingual pretrained representations are suboptimal for multi-lingual transfer.
Approach:	They propose a method that enforces consistency between standard and probabilistic segmentations.
Outcome:	The proposed method improves the effectiveness of cross-lingual transfer by 2.5 points over standard methods.

Copied to clipboard

Xinyi Wang, Yulia Tsvetkov, Sebastian Ruder, Graham Neubig

RACQC: Advanced Retrieval-Augmented Generation for Chinese Query Correction (2025.findings-emnlp)

Challenge:	Specialized language and task adapters have been proposed to facilitate cross-lingual transfer of multilingual pretrained models.
Approach:	They propose a method that optimizes the ensemble weights of pretrained adapters for each test sentence by minimizing the entropy of its predictions.
Outcome:	The proposed method improves robustness to uncovered languages without training new adapters.

Copied to clipboard

Jinbo Su, Lingzhe Gao, Wei Li, Shihao Liu, Haojie Lei, Xinyi Wang, Yuanzhao Guo, Ke Wang, Daiting Shi, Dawei Yin

Information Retrieval

Challenge:	Large language models (LLMs) exhibit remarkable capabilities across many tasks, but face critical challenges in the CSC scenario: (1) poor generalization to rare entities in open-domain searches; and (2) failure to adapt to temporal entity variations due to static parameters, resulting in serious over-correction issues.
Approach:	They propose a Chinese Spelling Check system with RAG and multi-task learning that integrates dynamic knowledge retrieval and entity-centric RAG to address rare entities.
Outcome:	The proposed system outperforms existing baselines in the CSC task and achieves a maximum improvement of +9.92% on the search scenario benchmark and +3.2% on the general-domain dataset.

ReflectEvo: Improving Meta Introspection of Small LLMs by Learning Self-Reflection (2025.findings-acl)

Copied to clipboard

Jiaqi Li, Xinyi Dong, Yang Liu, Zhizhuo Yang, Quansen Wang, Xiaobo Wang, Song-Chun Zhu, Zixia Jia, Zilong Zheng

Reasoning

Challenge:	ReflectEvo-460k is a large-scale, comprehensive, self-generated reflection dataset with broadened instructions and diverse multi-domain tasks.
Approach:	They propose a pipeline that iteratively generates self-reflection for self-training and a large-scale reflection dataset with broadened instructions and diverse multi-domain tasks.
Outcome:	The proposed pipeline improves Llama-3 reasoning ability by up to 71.2% and Mistral by upto 44.4%.

MTP-RL: Acceleration of Reinforcement Learning Rollouts with Policy-Aligned Multi-Token Prediction (2026.findings-acl)

Copied to clipboard

Ke Wang, Aohan Zeng, Zhengxiao Du, Yuxuan Hu, Bohan Zhang, Xinyi Wang, Jie Tang, Jing Zhang

Expanding Pretrained Models to Thousands More Languages via Lexicon-based Adaptation (2022.acl-long)

Challenge:	Reinforcement learning (RL) is widely applied to boost the performance of pretrained models, yet its training efficiency is severely constrained by rollout generation.
Approach:	They propose a framework that accelerates the rollout phase for diverse models by equipping a pipeline to equip the multi-layer parameter-sharing MTP for all models and an advantage-aware MTP optimization strategy.
Outcome:	The proposed framework achieves stable growth of acceptance length during RL training, and also accelerates RL rollouts, achieving an average 23.1%–55.3% reduction in rollout time compared to baselines.

Copied to clipboard

Xinyi Wang, Sebastian Ruder, Graham Neubig

RouteMoA: Dynamic Routing without Pre-Inference Boosts Efficient Mixture-of-Agents (2026.acl-long)

Challenge:	Recent studies have found that the performance of multilingual pretrained models is highly dependent on the availability of monolingual or parallel text in a target language.
Approach:	They propose to use bilingual lexicons to synthesize textual or labeled data and combine it with monolingual or parallel text when available.
Outcome:	The proposed methods improve performance for 19 under-represented languages with and without extra monolingual text.

Copied to clipboard

Jize Wang, Han Wu, Zhiyuan You, Yiming Song, Yijun Wang, Zifei Shan, Yining Li, Songyang Zhang, Xinyi Le, Cailian Chen, Xinping Guan, Dacheng Tao

Green & Sustainable NLP

Challenge:	Existing methods for mixing-of-agents (MoA) lack model selection criteria and struggle with large model pools.
Approach:	They propose a mixture-of-agents framework with dynamic routing that uses a lightweight scorer to perform initial screening and refines the model scores through self- and cross-assessment.
Outcome:	The proposed framework outperforms existing methods for large model pools and tasks . it reduces cost by 89.8% and latency by 63.6% in the large-scale model pool.

Continual Event Extraction with Semantic Confusion Rectification (2023.emnlp-main)

Copied to clipboard

Zitao Wang, Xinyi Wang, Wei Hu

CoIR: A Comprehensive Benchmark for Code Information Retrieval Models (2025.acl-long)

Challenge:	Existing studies focus on continual event extraction to extract incessantly emerging information . the semantic confusion on event types stems from the annotations of the same text being updated over time .
Approach:	They propose a continual event extraction model with semantic confusion rectification to reduce semantic confusion.
Outcome:	The proposed model outperforms state-of-the-art models and is proficient in imbalanced datasets.

Copied to clipboard

Xiangyang Li, Kuicai Dong, Yi Quan Lee, Wei Xia, Hao Zhang, Xinyi Dai, Yasheng Wang, Ruiming Tang

Information Retrieval

Challenge:	Existing methods and benchmarks for information retrieval are inadequately representing the diversity of code in various domains and tasks.
Approach:	They propose a benchmark specifically designed to assess code retrieval capabilities.
Outcome:	The proposed benchmark aims to invigorate research in the code retrieval domain . it shares the same data schema as other popular benchmarks like MTEB and BEIR .

Boost, Disentangle, and Customize: A Robust System2-to-System1 Pipeline for Code Generation (2025.findings-acl)

Copied to clipboard

Kounianhua Du, Hanjing Wang, Jianxing Liu, Jizheng Chen, Xinyi Dai, Yasheng Wang, Ruiming Tang, Yong Yu, Jun Wang, Weinan Zhang

Discovering Semantic Subdimensions through Disentangled Conceptual Representations (2025.findings-emnlp)

Challenge:	Existing systems 2 methods for code generation are difficult to implement due to the complex hidden reasoning process and heterogeneous data distribution.
Approach:	They propose a framework that Boosts reasoning exploration via multi-agent collaboration and Disentangles heterogeneous data into specialized experts.
Outcome:	The proposed framework outperforms state-of-the-art methods on APPS and CodeContest benchmarks and achieves 73.8% accuracy on hard problems.

Copied to clipboard

Yunhao Zhang, Shaonan Wang, Nan Lin, Xinyi Dong, Chong Li, Chengqing Zong

Explainability & Interpretability in NLP

Challenge:	Existing approaches focus on predefined dimensions that overlook finer conceptual distinctions . a new framework is proposed to investigate the subdimensions underlying coarse-grained semantic dimensions .
Approach:	They propose a framework that decomposes word embeddings into multiple sub-embeddings . they propose to map these subdimensions to brain activation to assess their plausibility .
Outcome:	The proposed framework decomposes word embeddings from large language models into sub-embeddings, each encoding specific semantic information.

Don’t Tell the Answer, Truly Guide the Reasoning During RL Rollouts (2026.findings-acl)

Copied to clipboard

Xinyi Wang, Jinyi Han, Zishang Jiang, Tingyun li, Jiaqing Liang, Sihang Jiang, Zhaoqian Dai, Ma Shuguang, Fei Yu, Yanghua Xiao

Reasoning

Challenge:	Existing methods such as GRPO often break down when task difficulty exceeds the model’s capacity, resulting in sparse rewards and inefficient training.
Approach:	They propose to measure the compatibility between external guidance and a model's intrinsic policy by introducing an adaptive framework to enhance reasoning performance while explicitly preserving high Affinity.
Outcome:	The proposed framework outperforms baseline models while maintaining high Affinity.

ADaPT: Token-Level Decoupling for Efficient Large Reasoning Models (2026.findings-acl)

Copied to clipboard

Tingyun li, Zishang Jiang, Jinyi Han, Xinyi Wang, Sihang Jiang, Han Xia, Zhaoqian Dai, Ma Shuguang, Fei Yu, Jiaqing Liang, Yanghua Xiao

Green & Sustainable NLP

Challenge:	Existing efficiency-oriented methods attempt to shorten or mix reasoning strategies, yet often degrade reasoning capability.
Approach:	They propose a token-level dual-process framework that explicitly decouples efficiency and correctness signals during training.
Outcome:	The proposed framework reduces inference cost while maintaining strong reasoning ability across multiple benchmarks.

mmT5: Modular Multilingual Pre-Training Solves Source Language Hallucinations (2023.findings-emnlp)

Copied to clipboard

Jonas Pfeiffer, Francesco Piccinno, Massimo Nicosia, Xinyi Wang, Machel Reid, Sebastian Ruder

RethinkMCTS: Refining Erroneous Thoughts in Monte Carlo Tree Search for Code Generation (2025.emnlp-main)

Challenge:	Recent large language models display surprising multilingual capabilities despite being pre-trained on English data.
Approach:	They propose a multilingual sequence-to-sequence model that disentangles language-specific information from language-agnostic information.
Outcome:	The proposed model outperforms existing models on representative natural language understanding and generation tasks in 40+ languages.

Copied to clipboard

Qingyao Li, Wei Xia, Xinyi Dai, Kounianhua Du, Weiwen Liu, Yasheng Wang, Ruiming Tang, Yong Yu, Weinan Zhang

Domain Differential Adaptation for Neural Machine Translation (D19-56)

Challenge:	Existing tree search methods neglect the underlying reasoning process, resulting in poor search quality.
Approach:	They propose a framework that systematically explores and refines the reasoning process for code generation by using a tree search engine and a reflection mechanism.
Outcome:	The proposed framework outperforms existing methods in the code generation domain.

Copied to clipboard

Zi-Yi Dou, Xinyi Wang, Junjie Hu, Graham Neubig

Machine Translation

Challenge:	Neural networks are data hungry and domain sensitive, so it is difficult to obtain labeled data for every domain.
Approach:	They propose a framework for domain adaptation where we model the difference between domains instead of smoothing over them.
Outcome:	The proposed framework improves on domain adaptation in multiple experimental settings.

A Multi-dimensional Evaluation of Tokenizer-free Multilingual Pretrained Models (2023.findings-eacl)

Copied to clipboard

Jimin Sun, Patrick Fernandes, Xinyi Wang, Graham Neubig

Syntactic Text Processing

Challenge:	Recent work on tokenizer-free models shows promising results in cross-lingual transfer . previous work focused on reporting accuracy on a limited set of tasks and data settings .
Approach:	They compare tokenizer-free and subword-based models using various dimensions . they find subword models are still the most practical choice in many settings .
Outcome:	The proposed model improves cross-lingual transfer and reduces engineering overhead.

JanusMM: A Benchmark for Self-Deprecation Understanding in Real-World Multimodal Conversations (2026.acl-long)

Copied to clipboard

Xinyi Xu, Bingguang Hao, Yongyi Xiong, Zimo Chen, Xinchen Liu, Hongxin Guo, Xuelong Wang, Silin Zhou, Shihan Dou

Dialogue Systems & Conversational Agents

Challenge:	Multimodal large language models (MLLMs) are a common communicative strategy in human society, often using image-text interplay to express emotions and intentions.
Approach:	They propose to evaluate multimodal large language models (MLLMs)' understanding of self-deprecation in real-world conversations using 2,016 bilingual memes.
Outcome:	The proposed framework evaluates MLLMs' understanding of self-deprecation in real-world conversations.

CodePRM: Execution Feedback-enhanced Process Reward Model for Code Generation (2025.findings-acl)

Copied to clipboard

Qingyao Li, Xinyi Dai, Xiangyang Li, Weinan Zhang, Yasheng Wang, Ruiming Tang, Yong Yu

compare-mt: A Tool for Holistic Comparison of Language Generation Systems (N19-4)

Challenge:	Recent advances in code generation focus on optimizing the thought process, but lack effective process supervision, making it difficult to optimize the thoughts.
Approach:	They propose a method that leverages the code execution feedback to build a code PRM by collecting a large dataset of thought traces and then training it to take both the reasoning process and code execution as input.
Outcome:	The proposed approach outperforms baselines and strong LLMs in the inference stage.

Copied to clipboard

Graham Neubig, Zi-Yi Dou, Junjie Hu, Paul Michel, Danish Pruthi, Xinyi Wang

Text Generation

Challenge:	Unlike machine translation, natural language outputs are nuanced and there are no clear yes/no distinctions about whether they are correct or not.
Approach:	They describe compare-mt, a tool for holistic analysis and comparison of the results of systems for language generation tasks such as machine translation.
Outcome:	The compare-mt tool is an open-source pure-python package that has already proven useful to generate analyses that have been used in our papers.

Modeling Disclosive Transparency in NLP Application Descriptions (2021.emnlp-main)

Copied to clipboard

Michael Saxon, Sharon Levy, Xinyi Wang, Alon Albalak, William Yang Wang

Challenge:	Broader disclosive transparency is difficult to define and quantify, authors say . previous work has demonstrated trade-offs and negative consequences to disclosing transparency .
Approach:	They propose to use neural language model-based probabilistic metrics to model disclosive transparency . they demonstrate that they correlate with user and expert opinions of system transparency a valid objective proxy .
Outcome:	The proposed metrics correlate with user and expert opinions of system transparency, making them a valid objective proxy.

Learning Mutually Informed Representations for Characters and Subwords (2024.findings-naacl)

Copied to clipboard

Yilin Wang, Xinyi Hu, Matthew Gormley

Mem2Evolve: Towards Self-Evolving Agents via Co-Evolutionary Capability Expansion and Experience Distillation (2026.acl-long)

Challenge:	Pretrained language models rely on subword tokenization to process text as a sequence of subwords.
Approach:	They propose a character-subword language model that integrates character and subword modalities into one model.
Outcome:	The proposed model outperforms its backbone language models on English sequence labeling and classification tasks.

Copied to clipboard

Zihao Cheng, Zeming Liu, Yingyu Shan, Xinyi Wang, Xiangrong Zhu, Yunpu Ma, Hongru Wang, Yuhang Guo, Wei Lin, Yunhong Wang

XTREME-UP: A User-Centric Scarce-Data Benchmark for Under-Represented Languages (2023.findings-emnlp)

Challenge:	Existing frameworks that focus on static tools and static assets are ineffective for self-evolving agents.
Approach:	They propose a paradigm of co-evolutionary Capability Expansion and Experience Distillation that leverages accumulated experience to guide dynamic creation of assets.
Outcome:	The proposed framework improves performance in single-task and cross-task settings by 18.53% over standard LLMs, 11.80% over agents evolving solely through experience, and 6.46% over those evolving solelly through asset creation.

Copied to clipboard

Sebastian Ruder, Jonathan Clark, Alexander Gutkin, Mihir Kale, Min Ma, Massimo Nicosia, Shruti Rijhwani, Parker Riley, Jean-Michel Sarr, Xinyi Wang, John Wieting, Nitish Gupta, Anna Katanova, Christo Kirov, Dana Dickinson, Brian Roark, Bidisha Samanta, Connie Tao, David Adelani, Vera Axelrod, Isaac Caswell, Colin Cherry, Dan Garrette, Reeve Ingle, Melvin Johnson, Dmitry Panteleev, Partha Talukdar

Mixture of LoRA Experts for Continual Information Extraction with LLMs (2025.findings-emnlp)

Challenge:	Existing datasets are often informed by established research directions in the NLP community.
Approach:	They propose a benchmark to evaluate the capabilities of language models across 88 under-represented languages over 9 key user-centric technologies including ASR, OCR, MT, and information access tasks.
Outcome:	The proposed benchmark evaluates the capabilities of language models across 88 under-represented languages over 9 key user-centric technologies including ASR, OCR, MT, and information access tasks.

Copied to clipboard

Zitao Wang, Xinyi Wang, Wei Hu

Collaborative Generative AI: Integrating GPT-k for Efficient Editing in Text-to-Image Generation (2023.emnlp-main)

Challenge:	Existing methods to continual information extraction are either task-specialized for a single task or suffer from catastrophic forgetting and insufficient knowledge transfer in continual IE.
Approach:	They propose a continual IE model that uses token-level mixture of LoRA experts with LLMs to extract emerging information across diverse IE tasks incessantly while avoiding forgetting.
Outcome:	The proposed model achieves state-of-the-art performance, effectively mitigating catastrophic forgetting and enhancing knowledge transfer in continual IE.

Copied to clipboard

Wanrong Zhu, Xinyi Wang, Yujie Lu, Tsu-Jui Fu, Xin Wang, Miguel Eckstein, William Wang

Me-Agent: A Personalized Mobile Agent with Two-Level User Habit Learning for Enhanced Interaction (2026.findings-acl)

Challenge:	Experimental results show that GPT-k models focus more on inserting modifiers than predicting spontaneous changes in the primary subject matter.
Approach:	They compare the common edits made by humans and GPT-k models to examine their performance in prompting T2I.
Outcome:	The proposed models improve the prompt editing process by 20-30%, the authors show . they show that humans tend to replace words and phrases with modifiers .

Copied to clipboard

Shuoxin Wang, Chang Liu, Gowen Loo, Lifan Zheng, Kaiwen Wei, Huanqian Yan, Xinyi Zeng, Jingyuan Zhang, Yu Tian

DCMKC: A Dual Consistency Matching Approach for Multi-hop Question Answering in LLMs (2025.findings-emnlp)

Challenge:	Existing Large Language Model (LLM)-based mobile agents follow explicit user instructions without personalized needs.
Approach:	They propose a user preference learning strategy enhanced with a Personal Reward Model to improve personalization performance.
Outcome:	The proposed agent achieves state-of-the-art performance while maintaining competitive instruction execution performance.

Copied to clipboard

Xinyi Wang, Yiping Song, Chang Liu, Tingjin Luo, Bo Liu, Zheng Xie, Minlie Huang

Structured Data in NLP

Challenge:	Existing reasoning based on chains of thought (CoTs) fails to find logical connections between reasoning steps .
Approach:	They propose a method to match KG reasoning chains with CoTs based on semantic similarity . they use a knowledge graph to find relevant information "within" each reasoning step .
Outcome:	The proposed method outperforms baselines on multi-answer questions with 5.1% improvement over baselines.

Improving Target-side Lexical Transfer in Multilingual Neural Machine Translation (2020.findings-emnlp)

Copied to clipboard

Luyu Gao, Xinyi Wang, Graham Neubig

SaCa: A Highly Compatible Reinforcing Framework for Knowledge Graph Embedding via Structural Pattern Contrast (2025.findings-emnlp)

Challenge:	Multilingual data is more beneficial for NMT models that translate from the LRL to a target language than those that translate into the LLLs.
Approach:	They propose a decoder that embeds character n-grams into NMT models that translate from an LRL to a target language.
Outcome:	The proposed decoder improves the performance of NMT models that translate from an LRL to a target language.

Copied to clipboard

Jiashi Lin, Changhong Jiang, Yixiao Wang, Xinyi Zhu, Zhongtian Hu, Wei Zhang

Exploring Fine-Grained Human Motion Video Captioning (2025.coling-main)

Challenge:	Existing knowledge Graph Embedding approaches lack structural semantics of knowledge graphs . structure-aware calibration (SaCa) is a framework designed to calibrate KGEs based on global structural patterns.
Approach:	a new framework is designed to calibrate knowledge graphs using global structural patterns.
Outcome:	a new framework can calibrate KGE models using global structural patterns . the framework consistently boosts performance across ten models on link prediction and entity classification tasks .

Copied to clipboard

Bingchan Zhao, Xinyi Liu, Zhuocheng Yu, Tongchen Yang, Yifan Song, Mingyu Jin, Sujian Li, Yizhou Wang

Logic-LM: Empowering Large Language Models with Symbolic Solvers for Faithful Logical Reasoning (2023.findings-emnlp)

Challenge:	Existing video captioning models fail to capture nuanced semantics of videos . existing models generate coarse descriptions of human motions, resulting in poor quality .
Approach:	They construct a fine-grained human motion video captioning dataset named BoFiT and a model that generates fine-grain descriptions of human motions via prompting.
Outcome:	The proposed model outperforms existing models on comprehensive metrics.

Copied to clipboard

Liangming Pan, Alon Albalak, Xinyi Wang, William Wang

LRBench and Judge-R1: Principled Evaluation and Training of LLM-Based Judges for Long-Context Reasoning (2026.findings-acl)

Challenge:	Large Language Models (LLMs) have shown human-like reasoning abilities but struggle with complex logical problems.
Approach:	They propose a framework which integrates large language models with symbolic solvers to improve logical problem-solving by combining them with a self-refinement module.
Outcome:	The proposed framework improves logical problem-solving by integrating LLMs with symbolic solvers.

Copied to clipboard

Xinyi Zhao, Haoqi Hu, Ziyu Wang, Jinfeng Xiao

Reasoning

Challenge:	Existing benchmarks for evaluating large language models (LLMs) under long contexts are underexplored.
Approach:	They propose a large-scale benchmark for evaluating large language models (LLMs) that combines reinforcement learning with multi-turn search to enable grounded and principle-aware evaluation.
Outcome:	The proposed model outperforms single-turn baselines across domains and principles.

Serial Contrastive Knowledge Distillation for Continual Few-shot Relation Extraction (2023.findings-acl)

Copied to clipboard

Xinyi Wang, Zitao Wang, Wei Hu

PASUM: A Pre-training Architecture for Social Media User Modeling Based on Text Graph (2024.lrec-main)

Challenge:	Existing models for few-shot relation extraction (RE) are not suitable for continual few-sshot RE.
Approach:	They propose a new model to train a model for new relations with few labeled training data.
Outcome:	The proposed model outperforms state-of-the-art models on two benchmark datasets.

Copied to clipboard

Kun Wu, Xinyi Mou, Lanqing Xue, Zhenzhe Ying, Weiqiang Wang, Qi Zhang, Xuanjing Huang, Zhongyu Wei

PECO: Examining Single Sentence Label Leakage in Natural Language Inference Datasets through Progressive Evaluation of Cluster Outliers (2023.eacl-main)

Challenge:	Existing studies have incorporated different digital traces to better learn the representations of social media users, limited by overloaded text information and hard-to-collect social network information.
Approach:	They propose a Pre-training Architecture for Social Media User Modeling based on Text Graph and combine microblogs to represent social media users based upon the text graph model.
Outcome:	The proposed framework can represent users based on text even without social network information on microblogs.

Copied to clipboard

Michael Saxon, Xinyi Wang, Wenda Xu, William Yang Wang

ToolGate: Contract-Grounded and Verified Tool Execution for LLMs (2026.findings-acl)

Challenge:	Efforts to debias NLI have led to datasets that exhibit different kinds of bias than those shown before.
Approach:	They propose a new technique to detect and reduce single sentence label leakage . leakage is a problem with many modern NLI datasets, they argue . future work must prioritize reducing this problem, they write .
Outcome:	a new model-driven technique can detect leakage and detect subpopulations in the datasets which exhibit it . the proposed technique is based on the progressive evaluation of cluster outliers (PECO) . it allows objective measurement of leakage, and automatic detection of subpopulations in the data which exhibit leakage.

Copied to clipboard

Yanming Liu, Xinyue Peng, Jiannan Cao, Xinyi Wang, Songhang Deng, Jintao Chen, Jianwei Yin, Xuhong Zhang

GUI0: Self-Evolving Foundational GUI Agents in Super App Ecosystems (2026.acl-long)

Challenge:	Existing frameworks for tool-augmented LLMs rely heavily on natural language reasoning to determine when tools can be invoked and whether their results should be trusted.
Approach:	They propose a forward execution framework that provides logical safety guarantees and verifiable state evolution for LLM tool calling.
Outcome:	The proposed framework improves the reliability and verifiability of tool-augmented LLM systems while maintaining competitive performance on multi-step reasoning tasks.

Copied to clipboard

Xinyi Wang, Wei Dai, Kyle Qiao, Ke Wang, Peng Chen, Gang Cao, null Kangqin, Zhongpu Wang, Xiaode Zhang, Yanming Liu, Jihao Gu, Jingtao Xu, Gong Zhi

LEDOM: Reverse Language Model (2026.acl-long)

Challenge:	Automated interaction with graphical user interfaces (GUIs) is central to general artificial intelligence, but remains challenging within Super App ecosystems.
Approach:	They propose a framework synergizing autonomous data synthesis with dual-agent co-evolution . GUI0 establishes a domain-aware foundation model via synthesized corpora and employs curriculum-driven reinforcement learning .
Outcome:	The proposed framework outperforms Gemini-2.5-Pro and Claude-4-Sonnet in the SuperAPP benchmark and has universal efficacy across base models.

Copied to clipboard

Xunjian Yin, Sitao Cheng, Yuxi Xie, Xinyu Hu, Li Lin, Xinyi Wang, Liangming Pan, William Yang Wang, Xiaojun Wan

Target Conditioned Sampling: Optimizing Data Selection for Multilingual Neural Machine Translation (P19-1)

Challenge:	Autoregressive language models are trained exclusively left-to-right, yet they are limited in their ability to factorize text.
Approach:	They propose a purely reverse autoregressive language model that factorizes text as a product of left-to-right conditionals.
Outcome:	The proposed model can be used to score forward outputs using reverse posterior estimates.

Copied to clipboard

Xinyi Wang, Graham Neubig

MDIT-Bench: Evaluating the Dual-Implicit Toxicity in Large Multimodal Models (2025.findings-acl)

Challenge:	Existing studies show that training on a single related language is more effective than using all data.
Approach:	They propose an efficient algorithm that first samples a target sentence, and then conditionally samples its source sentence.
Outcome:	The proposed algorithm brings significant gains on three of four languages with minimal training overhead.

Copied to clipboard

Bohan Jin, Shuhan Qi, Kehai Chen, Xinyi Guo, Xuan Wang

Has It All Been Solved? Open NLP Research Questions Not Solved by Large Language Models (2024.lrec-main)

Challenge:	Large Multimodal Models (LMMs) have raised concerns about model toxicity.
Approach:	They propose a model to measure the toxicity gap between models and their hard level to determine whether they can handle dual-implicit toxicity.
Outcome:	The proposed model can handle dual-implicit toxicity effectively on 13 prominent LMMs, but its performance drops significantly in hard level.

Copied to clipboard

Oana Ignat, Zhijing Jin, Artem Abzaliev, Laura Biester, Santiago Castro, Naihao Deng, Xinyi Gao, Aylin Ece Gunal, Jacky He, Ashkan Kazemi, Muhammad Khalifa, Namho Koh, Andrew Lee, Siyang Liu, Do June Min, Shinka Mori, Joan C. Nwatu, Veronica Perez-Rosas, Siqi Shen, Zekun Wang, Winston Wu, Rada Mihalcea

Instruction-Tuning Data Synthesis from Scratch via Web Reconstruction (2025.findings-acl)

Challenge:	Recent advances in large language models have led to misleading public discourse that “it’s all been solved.”
Approach:	They identify 14 research areas encompassing 45 research directions that require new research and are not directly solvable by LLMs.
Outcome:	The research areas identified are 45 research directions that require new research and are not directly solvable by LLMs.

Copied to clipboard

Yuxin Jiang, Yufei Wang, Chuhan Wu, Xinyi Dai, Yan Xu, Weinan Gan, Yasheng Wang, Xin Jiang, Lifeng Shang, Ruiming Tang, Wei Wang

Low-Resource NLP

Challenge:	Existing methods for generating and curating high-quality instruction-tuning data rely heavily on the quality of seed data or strong assumptions about the structure and content of web documents.
Approach:	They propose a fully automated framework for synthesizing high-quality instruction-tuning (IT) data directly from raw web documents with minimal assumptions.
Outcome:	The proposed framework outperforms state-of-the-art baselines by 16.65% across four instruction-following benchmarks.

LLM×MapReduce: Simplified Long-Sequence Processing using Large Language Models (2025.acl-long)

Copied to clipboard

Zihan Zhou, Chong Li, Xinyi Chen, Shuo Wang, Yu Chao, Zhili Li, Haoyu Wang, Qi Shi, Zhixing Tan, Xu Han, Xiaodong Shi, Zhiyuan Liu, Maosong Sun

Adaptive Hinge Balance Loss for Document-Level Relation Extraction (2023.findings-emnlp)

Challenge:	Existing studies have focused on extending the context length of large language models (LLMs) due to their quadratic computational complexity and a lack of high-quality long training examples, most LLMs are trained with a limited window size.
Approach:	They propose a training-free framework that enables large language models to effectively process long texts using a divide-and-conquer strategy for comprehensive document understanding.
Outcome:	The proposed framework outperforms open-source and commercial long-context LLMs and is compatible with several models.

Copied to clipboard

Jize Wang, Xinyi Le, Xiaodi Peng, Cailian Chen

Position Paper: Data-Centric AI in the Age of Large Language Models (2024.findings-emnlp)

Challenge:	Document-Level Relation Extraction is a multi-label classification task . however, most entity pairs do not express any relations .
Approach:	They propose to use a distance between the classification threshold and predicted score to reduce the imbalance problem by balancing the easy negatives.
Outcome:	The proposed method is superior to other methods, the authors show . it reduces the threshold to an appropriate value, and increases quadratically with the number of entities.

Copied to clipboard

Xinyi Xu, Zhaoxuan Wu, Rui Qiao, Arun Verma, Yao Shu, Jingtan Wang, Xinyuan Niu, Zhenfeng He, Jiangwei Chen, Zijian Zhou, Gregory Kang Ruey Lau, Hieu Dao, Lucas Agussurja, Rachael Hwee Ling Sim, Xiaoqiang Lin, Wenyang Hu, Zhongxiang Dai, Pang Wei Koh, Bryan Kian Hsiang Low