REFINEMENT

Flexible Thinking for Multimodal Emotional Support Conversation via Reinforcement Learning (2025.findings-emnlp)

Challenge:	Large-scale code generation models such as Copilot and CodeT5 are expensive to train and re-train.
Approach:	They propose a benchmark for Continual Learning (CL) that covers a wide range of tasks with different input and output programming languages.
Outcome:	The proposed method improves on Prompt Pooling with Teacher Forcing, which suffers catastrophic forgetting due to stark distribution shifts in coding tasks.

Copied to clipboard

Fanfan Wang, Xiangqing Shen, Jianfei Yu, Rui Xia

Dialogue Systems & Conversational Agents

Challenge:	Current Chain-of-Thought based ESC methods often employ rigid, text-only reasoning, limiting adaptability in dynamic, multimodal interactions and introducing reasoning noise that degrades support quality.
Approach:	They propose a framework that integrates supervised fine-tuning with reinforcement learning to improve ESC models' response quality.
Outcome:	The proposed framework enables models to select contextually relevant thinking aspects: Visual Scene, Emotion, Situation, and Response Strategy.

DIAGRAMS : A Review Framework for Reasoning-Level Attribution in Diagram QA (2026.acl-demo)

Copied to clipboard

Anirudh Iyengar Kaniyar Narayana Iyengar, Tampu Ravi Kumar, Manan Suri, Raviteja Bommireddy, Dinesh Manocha, Puneet Mathur, Vivek Gupta

Question Answering

Challenge:	Diagram question answering (Diagram QA) requires reasoning-level attribution that links each question-answer pair to all visual regions needed to derive the answer.
Approach:	They propose a diagram question-answer review framework that decouples interface logic from dataset-specific JSON structures through an internal meta-schema and dataset adapters.
Outcome:	The proposed framework achieves 85.39% precision and 75.30% recall across six diagram QA datasets.

Critic-CoT: Boosting the Reasoning Abilities of Large Language Model via Chain-of-Thought Critic (2025.findings-acl)

Copied to clipboard

Xin Zheng, Jie Lou, Boxi Cao, Xueru Wen, Yuqiu Ji, Hongyu Lin, Yaojie Lu, Xianpei Han, Debing Zhang, Le Sun

Autoformalization in the Wild: Assessing LLMs on Real-World Mathematical Definitions (2025.emnlp-main)

Challenge:	Existing approaches to improve the reasoning performance of large language models rely on intuitive instance-level feedback, which limits the reasoning capabilities.
Approach:	They propose a framework that pushes LLMs toward System-2-like critic capability by using a step-wise CoT reasoning paradigm and automatic construction of weak-supervision data without human annotation.
Outcome:	The proposed model significantly improves task-solving performance by filtering out invalid solutions or iterative refinement.

Copied to clipboard

Lan Zhang, Marco Valentino, Andre Freitas

Numerical Reasoning

Challenge:	Large Language Models (LLMs) have demonstrated remarkable potential in assisting with mathematical reasoning on different downstream tasks.
Approach:	They propose two new tools for autoformalizing real-world mathematical definitions from Wikipedia and arXiv papers.
Outcome:	The proposed methods improve definitions by up to 16% and undefined errors by 43%.

Reasoning in Flux: Enhancing Large Language Models Reasoning through Uncertainty-aware Adaptive Guidance (2024.acl-long)

Copied to clipboard

Zhangyue Yin, Qiushi Sun, Qipeng Guo, Zhiyuan Zeng, Xiaonan Li, Junqi Dai, Qinyuan Cheng, Xuanjing Huang, Xipeng Qiu

MotivGraph-SoIQ: Integrating Motivational Knowledge Graphs and Socratic Dialogue for Enhanced LLM Ideation (2025.findings-emnlp)

Challenge:	Extensive experiments across various reasoning tasks demonstrate that UAG not only enhances the reasoning abilities of LLMs but consistently outperforms several strong baselines with minimal computational overhead.
Approach:	They propose an approach to guide LLMs onto an accurate and reliable trajectory by identifying and adjusting uncertainty signals within each step of the reasoning chain.
Outcome:	The proposed approach outperforms strong baselines and outperformed strong models with minimal computational overhead.

Copied to clipboard

Xinping Lei, Tong Zhou, Yubo Chen, Kang Liu, Jun Zhao

Natural Language Interfaces

Challenge:	Large Language Models (LLMs) have limitations in grounding ideas and mitigating confirmation bias during refinement.
Approach:	They propose a framework that integrates a Motivational Knowledge Graph with a Q-Driven Socratic Ideator to enhance LLM ideation.
Outcome:	The proposed framework enhances LLM ideation by integrating a Motivational Knowledge Graph with a Q-Driven Socratic Ideator.

From Implicit Exploration to Structured Reasoning: Guideline and Refinement for LLMs (2025.findings-emnlp)

Copied to clipboard

Jiaxiang Chen, Zhuo Wang, Mingxi Zou, Zhucong Li, Zhijian Zhou, Song Wang, Zenglin Xu

Reasoning

Challenge:	Existing models rely on implicit exploration, which leads to unstable reasoning paths and lack of error correction.
Approach:	They propose a framework that shifts from implicit exploration to structured reasoning through guideline and refinement.
Outcome:	The proposed model outperforms strong baselines on the Big-Bench Hard benchmark.

VoiSeR: A New Benchmark for Voice-Based Search Refinement (2021.eacl-main)

Copied to clipboard

Simone Filice, Giuseppe Castellucci, Marcus Collins, Eugene Agichtein, Oleg Rokhlenko

Multimodality

Challenge:	a new study shows that voice-based search systems are challenging to support in the context of the user intent of voice searches . support for voice-driven search, exploration, and refinement is a fundamental aspect of voice assistants .
Approach:	They propose to use crowdsourcing to collect voice-based search refinements . they use 10,000 search refinement utterances to annotate a search intent .
Outcome:	The proposed dataset shows that voice-based search refinements can support most common tasks . the study shows that the proposed dataset can support research in conversational query understanding .

Inducing Generalizable and Interpretable Lexica (2022.findings-emnlp)

Copied to clipboard

Yilin Geng, Zetian Wu, Roshan Santhosh, Tejas Srivastava, Lyle Ungar, João Sedoc

Explainability & Interpretability in NLP

Challenge:	Lexica are widely used as generalizable language features to predict sentiment, emotions, mental health, and personality.
Approach:	They propose to induce lexica using context-oblivious and context-aware approaches and compare their performance using crowd-worker assessment.
Outcome:	The proposed models can be induced using context-oblivious and context-aware approaches and evaluate their quality using crowd-worker assessment.

Let’s Ask Again: Refine Network for Automatic Question Generation (D19-1)

Copied to clipboard

Preksha Nema, Akash Kumar Mohankumar, Mitesh M. Khapra, Balaji Vasan Srinivasan, Balaraman Ravindran

Semantic Text Processing

Challenge:	Existing AQG models produce incomplete questions which look like incomplete drafts with scope for refinement.
Approach:	They propose a method which mimics the human process of generating questions by first creating an initial draft and then refining it.
Outcome:	The proposed method outperforms state-of-the-art methods on three datasets and improves on fluency and answerability metrics.

mmE5: Improving Multimodal Multilingual Embeddings via High-quality Synthetic Data (2025.findings-acl)

Copied to clipboard

Haonan Chen, Liang Wang, Nan Yang, Yutao Zhu, Ziliang Zhao, Furu Wei, Zhicheng Dou

Multimodality

Challenge:	Multimodal embedding models encode multimedia inputs into latent vector representations.
Approach:	They propose to synthesize multimodal multilingual data using a multimodal large language model . they identify three criteria for high-quality synthetic multimodal data .
Outcome:	The proposed model outperforms existing models on the MMEB Benchmark and the XTD benchmark.

On the Evaluation of Neural Selective Prediction Methods for Natural Language Processing (2023.acl-long)

Copied to clipboard

Zhengyao Gu, Mark Hopkins

Challenge:	Existing techniques for selective classification are lacking in the literature.
Approach:	They propose a methodological blueprint and a metric for calibrating confidence functions for selective prediction.
Outcome:	The proposed method improves on the GLUE benchmark and the proposed refinement metric provides a calibrated evaluation of confidence functions for selective prediction.

Prompt Chaining or Stepwise Prompt? Refinement in Text Summarization (2024.findings-acl)

Copied to clipboard

Shichao Sun, Ruifeng Yuan, Ziqiang Cao, Wenjie Li, Pengfei Liu

Low-Resource NLP

Challenge:	Large language models (LLMs) can improve summary quality by mirroring a human-like iterative process of critique and refinement starting from the initial draft.
Approach:	They propose to use Prompt Chaining and Stepwise Prompting to perform iterative refinement . they propose to combine the two methods to produce a more favorable outcome .
Outcome:	The proposed methods can improve summary quality by mirroring a human-like iterative process . the results show that the prompt chaining method produces a more favorable outcome .

Weakly-Supervised Concept-based Adversarial Learning for Cross-lingual Word Embeddings (D19-1)

Copied to clipboard

Haozhou Wang, James Henderson, Paola Merlo

Compartmentalised Agentic Reasoning for Clinical NLI (2026.findings-acl)

Challenge:	Current unsupervised adversarial methods for mapping monolingual word embeddings require high quality parallel data, such as a dictionary or sentence-aligned corpus.
Approach:	They propose a weakly-supervised adversarial training method which improves performance of previous unsupervised adversaries for most languages.
Outcome:	The proposed method improves the performance of previous unsupervised adversarial methods for most languages and especially for typologically distant language pairs.

Copied to clipboard

Mael Jullien, Andre Freitas, Marco Valentino, Lei Xu

Reasoning

Challenge:	Large language models produce fluent judgments for clinical natural language inference, yet fail when the decision requires the correct inferential schema rather than surface matching.
Approach:	They propose a compartmentalised agentic framework that routes each premise–statement pair to a reasoning family and applies a specialised solver with explicit verification and targeted refinement.
Outcome:	The proposed framework improves mean accuracy from 23% with direct prompting to 57%, with the largest gains on structurally demanding reasoning types.

P-QuASAR: A Unified Probabilistic Framework for Holistic Patent Quality Assessment and Refinement (2026.findings-acl)

Copied to clipboard

Xinyuan Song, Ziyi Ni, Fred Yang, Bo Zhang, Yijin Wang, Jane Zhang

Structured Data in NLP

Challenge:	Existing methods for assessing patent quality rely on modular pipelines or generic detectors, resulting in fragmented decisions and limited integration across quality dimensions.
Approach:	They propose a probabilistic framework that represents patent specifications as Quality Graphs.
Outcome:	The proposed framework outperforms existing methods on 500 patents against seven baselines.

ExplaGraphs: An Explanation Graph Generation Task for Structured Commonsense Reasoning (2021.emnlp-main)

Copied to clipboard

Swarnadeep Saha, Prateek Yadav, Lisa Bauer, Mohit Bansal

Reasoning

Challenge:	Current commonsense-reasoning tasks are discriminative in nature, where a model answers a multiple-choice question for a certain context.
Approach:	They propose a generative task that generates a commonsense-augmented graph for stance prediction by using a create-verify-and-refine graph collection framework.
Outcome:	The proposed model is able to generate a graph that serves as non-trivial, complete, and unambiguous explanation for the predicted stance.

Sticker-TTS: Learn to Utilize Historical Experience with a Sticker-driven Test-Time Scaling Framework (2025.emnlp-main)

Copied to clipboard

Jie Chen, Jinhao Jiang, Yingqian Min, Zican Dong, Shijie Wang, Xin Zhao, Ji-Rong Wen

Learning to Refine with Fine-Grained Natural Language Feedback (2024.findings-emnlp)

Challenge:	Large reasoning models have exhibited strong performance on complex reasoning tasks, but current test-time scaling methods rely on redundant sampling and ignore historical experience utilization.
Approach:	They propose a test-time scaling framework that coordinates three collaborative LRMs to iteratively explore and refine solutions guided by historical attempts.
Outcome:	The proposed framework surpasses strong baselines on three mathematical reasoning benchmarks, including AIME-24, AIME-25, and OlymMATH.

Copied to clipboard

Manya Wadhwa, Xinyu Zhao, Junyi Jessy Li, Greg Durrett

Evaluating the Creativity of LLMs in Persian Literary Text Generation (2025.findings-emnlp)

Challenge:	Recent work has explored the capability of large language models to identify and correct errors in LLM-generated responses.
Approach:	They propose to combine refinement with feedback into three distinct competencies . step 1: Detect, Critique, Refine gives a fine-grained feedback about errors .
Outcome:	The proposed method outperforms existing refinement approaches and models not fine-tuned for factuality critiquing.

Copied to clipboard

Armin Tourajmehr, Mohammad Reza Modarres, Yadollah Yaghoobzadeh

Data Interpreter: An LLM Agent for Data Science (2025.findings-acl)

Challenge:	Prior research has focused primarily on English, with limited exploration of non-English literary traditions and without standardized methods for assessing creativity.
Approach:	They build a dataset of user-generated Persian literary spanning 20 diverse topics and assess model outputs along four creativity dimensions .
Outcome:	The proposed models generate Persian literary text enriched with culturally relevant expressions.

Copied to clipboard

Sirui Hong, Yizhang Lin, Bang Liu, Bangbang Liu, Binhao Wu, Ceyao Zhang, Danyang Li, Jiaqi Chen, Jiayi Zhang, Jinlin Wang, Li Zhang, Lingyao Zhang, Min Yang, Mingchen Zhuge, Taicheng Guo, Tuo Zhou, Wei Tao, Robert Tang, Xiangtao Lu, Xiawu Zheng, Xinbing Liang, Yaying Fei, Yuheng Cheng, Yongxin Ni, Zhibin Gou, Zongze Xu, Yuyu Luo, Chenglin Wu

QADYNAMICS: Training Dynamics-Driven Synthetic QA Diagnostic for Zero-Shot Commonsense Question Answering (2023.findings-emnlp)

Challenge:	Large Language Models (LLMs) excel in various domains but face challenges when applied to data science workflows due to their complex, multi-stage nature.
Approach:	They propose a hierarchical graph-based agent that represents complexity and a progressive strategy for step-by-step verification, refinement, and consistent context management.
Outcome:	The proposed agent surpasses state-of-the-art baselines on the MATH dataset and performs better on InfiAgent-DABench.

Copied to clipboard

Haochen Shi, Weiqi Wang, Tianqing Fang, Baixuan Xu, Wenxuan Ding, Xin Liu, Yangqiu Song

Semantic Text Processing

Challenge:	Existing approaches to QA fine-tune language models on QA pairs constructed from CommonSense Knowledge Bases (CSKBs) however, current QA synthesis protocols introduce noise from the CSKB and generate ungrammatical questions and false negative options, which impede the model’s ability to generalize.
Approach:	They propose a framework to analyze the training dynamics of each QA pair at both the question level and option level, discarding machine-detectable artifacts and mislabeled or false-negative options.
Outcome:	The proposed framework outperforms baseline approaches while using only 33% of the synthetic data.

SAIR-Comb : A Structure-Aware Iterative Refinement Framework for Combinatorics Autoformalization (2026.acl-long)

Copied to clipboard

Weijie Jiang, Gaolei He, Beibei Xiong, Jianlin Wang, Zhengfeng Yang

Syntactic Text Processing

Challenge:	Large language models (LLMs) have catalyzed advances in mathematical reasoning, propelling the development of automated theorem proving (ATP).
Approach:	They propose a Structure-Aware Iterative Refinement framework for Combinatorics powered by Lean 4 and LLMs.
Outcome:	The proposed framework achieves strong performance on the specialized CombiBench while remaining highly competitive on general-domain benchmarks.

MT-Eval: A Multi-Turn Capabilities Evaluation Benchmark for Large Language Models (2024.emnlp-main)

Copied to clipboard

Wai-Chung Kwan, Xingshan Zeng, Yuxin Jiang, Yufei Wang, Liangyou Li, Lifeng Shang, Xin Jiang, Qun Liu, Kam-Fai Wong

SOPL: A Sequential Optimal Learning Approach to Automated Prompt Engineering in Large Language Models (2025.findings-emnlp)

Challenge:	Existing evaluation frameworks focus on single-turn evaluations, overlooking the models’ capabilities in multi-turn interactions.
Approach:	They propose a benchmark to evaluate the multi-turn conversational abilities of large language models (LLMs) by analyzing human-LLM conversations and constructing multi-turned queries for each category using GPT-4.
Outcome:	The proposed model outperforms open-source models in multi-turn tasks while retaining and recalling historical information.

Copied to clipboard

Shuyang Wang, Somayeh Moazeni, Diego Klabjan

Simple Role Assignment is Extraordinarily Effective for Safety Alignment (2026.findings-acl)

Challenge:	Using automated prompt engineering to identify effective features is essential for large language models.
Approach:	They propose an optimal learning framework for automated prompt engineering for black-box models . feature-based method is used to express prompt templates, which broadens the search space .
Outcome:	The proposed learning framework outperforms benchmark strategies on instruction induction tasks with limited budgets.

Copied to clipboard

Zhou Ziheng, Jiakun Ding, Zhaowei Zhang, Ruosen Gao, Ying Nian Wu, Demetri Terzopoulos, Yipeng Kang, Fangwei Zhong, Junqi Wang

Challenge:	a new study proposes a role-conditioned pipeline for value alignment . principles alone are incomplete, and they provide little guidance on when and how a value applies in context.
Approach:	They propose a role-conditioned pipeline with role-based critics and a model-free approach that is based on role conditioning.
Outcome:	The proposed approach outperforms principle-based, Chain-of-Thought and other benchmarks.

Agent-as-Judge for Factual Summarization of Long Narratives (2025.emnlp-main)

Copied to clipboard

Yeonseok Jeong, Minsoo Kim, Seung-won Hwang, Byung-Hak Kim

Summarization

Challenge:	Large Language Models (LLMs) have demonstrated near-human performance in summarization tasks based on traditional metrics such as ROUGE and BERTScore . however, these metrics do not adequately capture critical aspects of summarizing quality, such as factual accuracy, especially for long narratives.
Approach:	They propose a framework that evaluates and refines factuality in narrative summarization by leveraging a Character Knowledge Graph extracted from input narrative.
Outcome:	The proposed framework evaluates factuality and provides actionable guidance for refinement.

ReflectDiffu: Reflect between Emotion-intent Contagion and Mimicry for Empathetic Response Generation via a RL-Diffusion Framework (2025.acl-long)

Copied to clipboard

Jiahao Yuan, Zixiang Di, Zhiqing Cui, Guisong Yang, Usman Naseem

mrCAD: Multimodal Communication to Refine Computer-aided Designs (2025.findings-emnlp)

Challenge:	Existing models for empathetic dialogue generation neglect the intricate interplay between emotion and intent, leading to suboptimal controllability of empathy.
Approach:	They propose a framework that integrates emotion contagion and intent mimicry to enhance empathetic response generation.
Outcome:	The proposed framework outperforms existing models in relevance, controllability, and informativeness.

Copied to clipboard

William P McCarthy, Saujas Vaduguru, Karl D.d. Willis, Justin Matejka, Judith E Fan, Daniel Fried, Yewen Pu

Multimodality

Challenge:	generative AI excels at creating artifacts in a single turn but can struggle to make precise refinements that match our design intent.
Approach:	They propose to use multi-turn interactions to iterate and refine computer-aided designs (CADs) they use text and drawing to communicate with each other over multiple rounds of interaction .
Outcome:	mrCAD consists of 6,082 communication games, 15,163 instruction-execution rounds, played between 1,092 pairs of humans.

MMRefine: Unveiling the Obstacles to Robust Refinement in Multimodal Large Language Models (2025.findings-acl)

Copied to clipboard

Gio Paik, Geewook Kim, Jinbae Im

Robustness in NLP

Challenge:	Recent advances have enabled MLLMs to tackle complex challenges such as mathematical reasoning and multimodal understanding.
Approach:	They propose a multimodal refinement benchmark to evaluate the refinement capabilities of Multimodal Large Language Models (MLLMs) the benchmark categorizes errors into six error types to highlight areas for improvement in effective reasoning enhancement.
Outcome:	The proposed framework evaluates the refinement capabilities of multimodal large language models across six scenarios.

MTSQL-R1: Towards Long-Horizon Multi-Turn Text-to-SQL via Agentic Training (2026.acl-long)

Copied to clipboard

Taicheng Guo, Hai Wang, Chaochun Liu, Mohsen Golalikhani, Xin Chen, Xiangliang Zhang, Chandan K. Reddy

Turning Failures into Value: Negative Experience Replay for RLVR via Confidence Gating and Boundary Failure Sampling (2026.acl-long)

Challenge:	Existing systems for multi-turn Text-to-SQL are limited to a short-horizon paradigm, generating a query per turn without execution, explicit verification, and refinement, which leads to non-executable or incoherent outputs.
Approach:	They propose to train an agentic training framework for long-horizon multi-turn Text-to-SQL that uses a Markov Decision Process to generate a query per turn without execution, explicit verification, and refinement.
Outcome:	Experiments on CoSQL and SParC show that MTSQL-R1 consistently outperforms strong baselines, highlighting the importance of environment-driven verification and memory-guided refinement for conversational semantic parsing.

Copied to clipboard

Jialiang Guo, Fucheng Xiong, Xu He, Haodong Zhao, Xingyang li, Ke Zeng, Xunliang Cai