Papers by Yun-Nung Chen
Semantically-Aligned Equation Generation for Solving and Reasoning Math Word Problems (N19-1)
Copied to clipboard
| Challenge: | Existing methods to solve math word problems require accurate natural language understanding to bridge texts and math expressions. |
| Approach: | They propose a neural approach to automatically solve math word problems by operating symbols according to their semantic meanings in texts. |
| Outcome: | The proposed model outperforms state-of-the-art models and the best non-retrieval-based models over 10% accuracy in a Math23K dataset. |
What Does This Word Mean? Explaining Contextualized Embeddings with Natural Language Definition (D19-1)
Copied to clipboard
| Challenge: | Contextualized word embeddings have boosted many NLP tasks compared with static word embeds. |
| Approach: | They propose a framework that can explain word meanings given contextualized word embeddings for better interpretation. |
| Outcome: | The proposed framework can explain word meanings given contextualized word embeddings for better interpretation. |
Efficient Unseen Language Adaptation for Multilingual Pre-Trained Language Models (2024.emnlp-main)
Copied to clipboard
| Challenge: | Multilingual pre-trained language models (mPLMs) have demonstrated notable effectiveness in zero-shot cross-lingual transfer tasks. |
| Approach: | They propose a method that uses soft-prompt tuning to tune for language adaptation . prompt tuning outperforms continuously trained baselines on two benchmarks . |
| Outcome: | The proposed approach outperforms baselines on two text classification benchmarks while utilizing 0.28% of tuned parameters. |
Natural Language Generation by Hierarchical Decoding with Linguistic Patterns (N18-2)
Copied to clipboard
| Challenge: | a common and mostly adopted method is the rule-based (or template-based) method for natural language generation. |
| Approach: | They propose a hierarchical decoding NLG model based on linguistic patterns in different levels. |
| Outcome: | The proposed method outperforms the traditional one with a smaller model size. |
Visualizing Dialogues: Enhancing Image Selection through Dialogue Understanding with Large Language Models (2024.findings-acl)
Copied to clipboard
| Challenge: | Existing methods for dialogue-to-image retrieval are constrained by pre-trained vision language models. |
| Approach: | They leverage the reasoning capabilities of large language models to predict potential features in images to be shared based on dialogue context. |
| Outcome: | The proposed method outperforms existing methods significantly in terms of Recall@k. |
What Do Position Embeddings Learn? An Empirical Study of Pre-Trained Language Model Positional Encoding (2020.emnlp-main)
Copied to clipboard
| Challenge: | Existing work on pre-trained Transformers has focused on learning the meaning of positions . Embedding the position information in the self-attention mechanism is also an indispensable factor in NLP . |
| Approach: | They propose to use feature-level analysis to examine pre-trained Transformers' position embeddings . they also use empirical experiments to determine the appropriate positional encoding function . |
| Outcome: | The results of the empirical study can guide future work to choose the appropriate positional encoding function for specific tasks. |
Lifelong Language Knowledge Distillation (2020.emnlp-main)
Copied to clipboard
| Challenge: | Existing methods to perform lifelong language learning (LLL) on stream of different tasks are challenging . Existing models face catastrophic forgetting problem, which can be mitigated by lifelong learning . |
| Approach: | They propose a method that can be easily applied to existing LLL architectures to mitigate degradation. |
| Outcome: | The proposed method improves state-of-the-art models and reduces degradation compared to multi-task models. |
LLMs are Biased Evaluators But Not Biased for Fact-Centric Retrieval Augmented Generation (2025.findings-acl)
Copied to clipboard
| Challenge: | Recent studies have shown that large language models (LLMs) exhibit significant biases in evaluation tasks, especially in preferentially rating and favoring self-generated content. |
| Approach: | They propose to simulate two critical phases of retrieval-augmented generation (RAG) frameworks where keyword extraction and factual accuracy take precedence over stylistic elements. |
| Outcome: | The proposed model emulates two critical phases of the retrieval-augmented generation framework. |
CLUSE: Cross-Lingual Unsupervised Sense Embeddings (D18-1)
Copied to clipboard
| Challenge: | Existing models for learning bilingual sense embeddings that encode semantics are weak for recognizing multi-sense word representations . elucidation of word embeddables is difficult because they do not allow a word to have different meanings in different contexts. |
| Approach: | They propose a sense induction and representation learning model that learns bilingual sense embeddings that align well in the vector space. |
| Outcome: | The proposed model shows that the learned embeddings are aligned well in the vector space. |
Dual Supervised Learning for Natural Language Understanding and Generation (P19-1)
Copied to clipboard
| Challenge: | Natural language understanding and natural language generation are important research topics in the NLP and dialogue fields. |
| Approach: | They propose a dual-supervised learning framework for natural language understanding and generation on top of dual supervised learning. |
| Outcome: | The proposed framework boosts the performance of both tasks simultaneously in the benchmark experiments. |
Compound AI Systems Optimization: A Survey of Methods, Challenges, and Future Directions (2025.emnlp-main)
Copied to clipboard
| Challenge: | Recent advances in large language models (LLMs) and AI systems have led to a paradigm shift in the design and optimization of complex workflows. |
| Approach: | They propose a systematic review of recent progress in optimizing compound AI systems . they formalize the notion of compound AI system optimization and classify existing methods along several key dimensions . |
| Outcome: | The proposed methods outperform existing methods in the field of compound AI and highlight open research challenges and future directions. |
VisTW: Benchmarking Vision-Language Models for Taiwanese Mandarin in Taiwan (2026.findings-acl)
Copied to clipboard
| Challenge: | Vision-Language Models (VLMs) struggle in Taiwanese Mandarin environments due to region-specific orthographic and cultural context. |
| Approach: | They propose a human-grounded purity penalty for character mixing under Taiwanese-Mandarin-style prompts. |
| Outcome: | The proposed model outperforms the strongest open-weight baseline by 22 percentage points on dialogue tasks. |
Tree Transformer: Integrating Tree Structures into Self-Attention (D19-1)
Copied to clipboard
| Challenge: | Existing work on hierarchical structure in neural networks has not captured human intuitions about hierarchic structures. |
| Approach: | They propose to add an extra constraint to attention heads of the bidirectional Transformer encoder to encourage attention heads to follow tree structures. |
| Outcome: | The proposed model improves language modeling and learning more explainable attention scores. |
None of the Above, Less of the Right Parallel Patterns in Human and LLM Performance on Multi-Choice Questions Answering (2025.findings-acl)
Copied to clipboard
| Challenge: | Multiple-choice exam questions with “None of the above” (NA) options have been extensively studied in educational testing . however, their impact on Large Language Models (LLMs) evaluation remains underexplored . |
| Approach: | They conduct systematic experiments with 28 LLMs on the MMLU benchmark to examine how NA options affect model performance and confidence calibration. |
| Outcome: | The results highlight important implications for benchmark design and raise questions about LLMs’ ability to handle uncertainty in real-world applications. |
Modeling Diagnostic Label Correlation for Automatic ICD Coding (2021.naacl-main)
Copied to clipboard
| Challenge: | Existing work built a binary prediction for each label independently, ignoring the dependencies between labels. |
| Approach: | They propose a framework to capture the label correlation and train a reranking estimator to rescore the probability of each label set candidate generated by a base predictor. |
| Outcome: | The proposed framework improves on the best-performing predictors on MIMIC datasets. |
Discriminative Deep Dyna-Q: Robust Planning for Dialogue Policy Learning (D18-1)
Copied to clipboard
| Challenge: | Existing approaches to improve the effectiveness and robustness of Deep Dyna-Q (DDQ) are based on a discriminator to control the quality of simulated experiences and to improve learning. |
| Approach: | They propose to use an RNN-based discriminator to control the quality of simulated experience to improve the effectiveness and robustness of Deep Dyna-Q. |
| Outcome: | The proposed framework outperforms DDQ by controlling the quality of simulated experience used for planning. |
Editing the Mind of Giants: An In-Depth Exploration of Pitfalls of Knowledge Editing in Large Language Models (2024.findings-emnlp)
Copied to clipboard
| Challenge: | Knowledge editing is a promising technique for updating factual knowledge in large language models (LLMs) but studies have identified side effects such as knowledge distortion and the deterioration of general abilities that have emerged after editing. |
| Approach: | They propose to evaluate the side effects of knowledge editing in large language models using metrics and benchmarks. |
| Outcome: | The results of the study highlight the limitations of current knowledge editing methods and outline potential research directions. |
Transferring Textual Preferences to Vision-Language Understanding through Model Merging (2025.acl-short)
Copied to clipboard
| Challenge: | Large vision-language models (LVLMs) perform outstandingly across multimodal tasks, but training them with preference data is computationally expensive. |
| Approach: | They propose to merge text-based reward models with LVLMs to create visionlanguage reward models (VLRMs) this approach offers an efficient method for incorporating textual preferences into LVRMs. |
| Outcome: | The proposed model improves over LVLMs’ scoring and text-based RMs, and offers an efficient method for incorporating textual preferences into LVRMs. |
FlowDelta: Modeling Flow Information Gain in Reasoning for Conversational Machine Comprehension (D19-58)
Copied to clipboard
| Challenge: | Existing machine comprehension models focus on a single-turn setting and do not account for previous reasoning processes. |
| Approach: | They propose to explicitly model the information gain through the dialogue reasoning . they propose to apply the proposed mechanism to other machine comprehension models . |
| Outcome: | The proposed model achieves state-of-the-art performance in a conversational QA dataset QuAC and a sequential instruction understanding dataset SCONE. |
Self-ICL: Zero-Shot In-Context Learning with Self-Generated Demonstrations (2023.emnlp-main)
Copied to clipboard
| Challenge: | Large language models (LLMs) have shown striking ability to adapt to target tasks with a few input-output demonstrations. |
| Approach: | They propose a framework which bootstraps LMs’ intrinsic capabilities to perform zero-shot ICL. |
| Outcome: | The proposed framework outperforms baselines on 23 BIG-Bench Hard tasks on average accuracy and head-to-head comparison. |
Deep Learning for Dialogue Systems (C18-3)
Copied to clipboard
| Challenge: | Using deep learning to build robust and scalable spoken dialogue systems is still a challenging task. |
| Approach: | tutorial focuses on an overview of dialogue system development . goal-oriented spoken dialogue systems are most prominent component in virtual personal assistants . |
| Outcome: | This tutorial focuses on an overview of dialogue system development while summarizing the challenges. |
FactAlign: Long-form Factuality Alignment of Large Language Models (2024.findings-emnlp)
Copied to clipboard
| Challenge: | Large language models have demonstrated significant potential as the next-generation information access engines, but reliability is hindered by issues of hallucination and generating non-factual content. |
| Approach: | They propose a novel alignment framework that enhances the factuality of LLMs’ long-form responses while maintaining their helpfulness. |
| Outcome: | The proposed framework improves factuality of LLMs while maintaining helpfulness. |
How Time Matters: Learning Time-Decay Attention for Contextual Spoken Language Understanding in Dialogues (N18-1)
Copied to clipboard
| Challenge: | Spoken language understanding (SLU) is an essential component in conversational systems. |
| Approach: | They propose a universal time-decay attention mechanism that can be used to decay utterances on the sentence-level and speaker-level. |
| Outcome: | The proposed model significantly improves the state-of-the-art model for contextual understanding performance on the benchmark Dialogue State Tracking Challenge (DSTC4) dataset. |
Let Me Speak Freely? A Study On The Impact Of Format Restrictions On Large Language Model Performance. (2024.emnlp-industry)
Copied to clipboard
| Challenge: | Structured generation is used to extract key output information from large language models (LLMs). |
| Approach: | They examine whether constraints on generation space impact LLMs’ abilities, including reasoning and domain knowledge comprehension. |
| Outcome: | The proposed model is based on a few-shot in-context learning and instruction-following capabilities. |
Unsupervised Multilingual Dense Retrieval via Generative Pseudo Labeling (2024.findings-eacl)
Copied to clipboard
| Challenge: | Existing sparse retrieval methods often yield inferior performance in multilingual retrieval, requiring a large amount of paired data, which is costly. |
| Approach: | They propose an Unsupervised Multilingual dense Retriever trained without paired data which iteratively improves performance of multilingual retrievers. |
| Outcome: | The proposed framework outperforms supervised baselines on two benchmark datasets and shows that iterative training improves the performance. |
Learning Spoken Language Representations with Neural Lattice Language Modeling (2020.acl-main)
Copied to clipboard
| Challenge: | Existing methods for pretraining language models do not consider spoken language properties. |
| Approach: | They propose a framework that trains neural lattice language models to provide contextualized representations for spoken language understanding tasks. |
| Outcome: | The proposed framework outperforms baselines on spoken inputs on intent detection and dialogue act recognition datasets. |
Leveraging Hierarchical Category Knowledge for Data-Imbalanced Multi-Label Diagnostic Text Understanding (D19-62)
Copied to clipboard
| Challenge: | Existing work on automatic diagnosis prediction has considered the output labels independently, so that the codes with few samples are difficult to learn. |
| Approach: | They propose to leverage domain knowledge to predict diagnostic codes given the descriptive present illness in electronic health records by leveraging hierarchical category knowledge. |
| Outcome: | The proposed model can efficiently utilize category knowledge and provide informative cues to improve the top-ranked diagnostic codes which is better than the prior state-of-the-art. |
Visually-Enhanced Phrase Understanding (2023.findings-acl)
Copied to clipboard
| Challenge: | Large-scale vision-language pre-training models generate high-quality textual representations, which often outperform models that are purely text-based, such as BERT. |
| Approach: | They propose to utilize both textual and visual encoders of multi-modal pre-trained models to enhance language understanding tasks by generating an image associated with a textual prompt. |
| Outcome: | The proposed method outperforms models that are purely text-based on visual and textual understanding tasks and significantly improves the entity clustering task. |
Zero-Shot Rationalization by Multi-Task Transfer Learning from Question Answering (2020.findings-emnlp)
Copied to clipboard
| Challenge: | Existing methods to extract rationales from input text are difficult and impractical. |
| Approach: | They propose a method that leverages multi-task learning and transfer learning to generate rationales through question answering in a zero-shot fashion. |
| Outcome: | The proposed method achieves comparable or even better performance without supervised signal for two benchmark rationalization datasets. |
Towards Unsupervised Language Understanding and Generation by Joint Dual Learning (2020.acl-main)
Copied to clipboard
| Challenge: | Existing work exploits dual property between understanding and generation to improve performance of modular dialogue systems. |
| Approach: | They propose a dual supervised learning framework that exploits the dual property between understanding and generation. |
| Outcome: | The proposed framework improves both NLU and NLG performance by incorporating supervised and unsupervised learning algorithms. |
Towards Understanding of Medical Randomized Controlled Trials by Conclusion Generation (D19-62)
Copied to clipboard
| Challenge: | Using machine learning to interpret large amounts of data can be over-whelming for clinicians. |
| Approach: | They propose to use PubMed 200k RCT sentence classification dataset to generate RCT conclusion generation task. |
| Outcome: | The proposed model improves quality and correctness in generated conclusions compared to baseline model . the proposed model is not suitable for all RCTs, but it could be improved . |
SalesBot: Transitioning from Chit-Chat to Task-Oriented Dialogues (2022.acl-long)
Copied to clipboard
| Challenge: | Until now, researchers have separated open-domain and task-oriented dialogues into two different types due to their different purposes. |
| Approach: | They propose a framework to automatically generate many dialogues without human involvement . the framework can be easily leveraged to generate unlimited dialogues in target scenarios . |
| Outcome: | The proposed framework can automatically generate many dialogues without human involvement . the human evaluation shows that the generated dialogues have a reasonable quality . |
Two Tales of Persona in LLMs: A Survey of Role-Playing and Personalization (2024.findings-emnlp)
Copied to clipboard
| Challenge: | Existing literature on leveraging persona in large language models is disorganized and lacks a systematic taxonomy . leveraging peopleas has resurfaced as an ideal lens for adapting LLMs for specific contexts . |
| Approach: | They propose to categorize current research on leveraging persona in large language models . they propose to use a comprehensive survey to categorize existing studies . |
| Outcome: | The proposed framework is a promising framework for tailoring large language models to specific contexts. |
Slot-Gated Modeling for Joint Slot Filling and Intent Prediction (N18-2)
Copied to clipboard
| Challenge: | Existing approaches for slot filling and intent detection have independent attention weights, but they suffer from error propagation due to their independent models. |
| Approach: | They propose a slot gate that focuses on learning the relationship between intent and slot attention vectors to obtain better semantic frame results by the global optimization. |
| Outcome: | The proposed model significantly improves sentence-level semantic frame accuracy with 4.2% and 1.9% relative improvement compared to the attentional model on benchmark ATIS and Snips datasets respectively. |
Injecting Salesperson’s Dialogue Strategies in Large Language Models with Chain-of-Thought Reasoning (2024.findings-acl)
Copied to clipboard
| Challenge: | Recent research in dialogue systems focuses on task-oriented (TOD) and open-domain (chit-chat) dialogues. |
| Approach: | They propose to use chit-chat to simulate task-oriented dialogues to train sales agents. |
| Outcome: | The proposed model improves coherence and reduces aggression, improving model learning for sales-customer interactions. |
Zero-Shot Prompting for Implicit Intent Prediction and Recommendation with Commonsense Reasoning (2023.findings-acl)
Copied to clipboard
| Challenge: | Currently, intelligent assistants require explicit user requests to perform tasks or services, leading to lengthy and complex conversations. |
| Approach: | They propose a framework that automatically infers implicit intents from user utterances and prompts a large pre-trained language model to suggest suitable task-oriented bots. |
| Outcome: | The proposed framework reduces interaction complexity and integrates domains and tasks. |
I Need Help! Evaluating LLM’s Ability to Ask for Users’ Support: A Case Study on Text-to-SQL Generation (2024.emnlp-main)
Copied to clipboard
| Challenge: | a new study examines the proactive ability of large language models to seek user support . without external feedback, many LLMs struggle to recognize their need for user support. |
| Approach: | They propose metrics to evaluate the trade-off between performance improvements and user burden . they also investigate whether LLMs can determine when to request user support . |
| Outcome: | The proposed metrics show that without external feedback, many LLMs struggle to recognize their need for user support. |
Pseudo2Real: Task Arithmetic for Pseudo-Label Correction in Automatic Speech Recognition (2026.findings-acl)
Copied to clipboard
| Challenge: | ASR models can be used to correct accent-specific errors without ground truth . pseudo-labels inherit the teacher model's systematic biases, authors say . |
| Approach: | They propose a parameter-space correction technique that captures pseudo-label biases . they propose achieving up to 35% relative WER reduction on a pseudo-labeled target model . |
| Outcome: | The proposed model achieves 35% relative WER reduction on ten African accents with the Whisper tiny model. |
Dual Inference for Improving Language Understanding and Generation (2020.findings-emnlp)
Copied to clipboard
| Challenge: | Existing studies have exploited the duality of the task pairs in machine translation and speech recognition. |
| Approach: | They propose to leverage the duality in the inference stage without retraining whole models. |
| Outcome: | The proposed method is effective in both NLU and NLG tasks, providing the great potential of practical use. |
QAInfomax: Learning Robust Question Answering System by Mutual Information Maximization (D19-1)
Copied to clipboard
| Challenge: | Existing models are not good at distinguishing distractor sentences which look related but do not answer the question. |
| Approach: | They propose a method to regularize question answering models by maximizing mutual information among passages, questions, and answers. |
| Outcome: | The proposed model achieves state-of-the-art on the Adversarial-SQuAD dataset. |
PairDistill: Pairwise Relevance Distillation for Dense Retrieval (2024.emnlp-main)
Copied to clipboard
| Challenge: | Recent advances in dense retrieval have demonstrated remarkable efficacy compared to traditional sparse retrieval methods. |
| Approach: | They propose to use pairwise relevance distillation to leverage pairwise reranking to enrich the training of dense retrieval models. |
| Outcome: | The proposed method outperforms existing methods and achieves state-of-the-art results on multiple benchmarks. |
Efficient Multi-Task Auxiliary Learning: Selecting Auxiliary Data by Feature Similarity (2021.emnlp-main)
Copied to clipboard
| Challenge: | Multi-task auxiliary learning uses a set of relevant auxiliary tasks to improve performance of a primary task. |
| Approach: | They propose a time-efficient sampling method to select the most beneficial sub-datasets from the auxiliary tasks to achieve efficient multi-task auxiliary learning. |
| Outcome: | The proposed method significantly outperforms random sampling and ST-DNN on three benchmark datasets. |
The Role of Exploration Modules in Small Language Models for Knowledge Graph Question Answering (2025.acl-srw)
Copied to clipboard
| Challenge: | Existing methods to integrate knowledge graphs into large language models often rely on proprietary or extremely large models . |
| Approach: | They propose to integrate knowledge graphs into reasoning processes of large language models . they propose to use simple and efficient exploration modules to handle knowledge graph traversal . |
| Outcome: | The proposed modules improve the performance of small language models on knowledge graph question answering tasks. |
DogeRM: Equipping Reward Models with Domain Knowledge through Model Merging (2024.emnlp-main)
Copied to clipboard
| Challenge: | Modern large language models (LLMs) showcase impressive capabilities across various tasks with aligning their behavior with human preferences. |
| Approach: | They propose a framework that integrates domain-specific knowledge into a general reward model by model merging. |
| Outcome: | The proposed framework improves performance across different benchmarks and provides detailed analysis showing the effects of model merging. |
Creativity in LLM-based Multi-Agent Systems: A Survey (2025.emnlp-main)
Copied to clipboard
Yi-Cheng Lin, Kang-Chieh Chen, Zhe-Yan Li, Tzu-Heng Wu, Tzu-Hsuan Wu, Kuan-Yu Chen, Hung-yi Lee, Yun-Nung Chen
| Challenge: | Large language model (LLM)-driven multi-agent systems (MAS) are transforming how humans and AIs collaboratively generate ideas and artifacts. |
| Approach: | They present a taxonomy of agent proactivity and persona design and an overview of generation techniques. |
| Outcome: | The proposed framework and roadmap offers a roadmap for advancing the development, evaluation, and standardization of creative MAS. |
DyKgChat: Benchmarking Dialogue Generation Grounding on Dynamic Knowledge Graphs (D19-1)
Copied to clipboard
| Challenge: | Existing work has not shown that knowledge-grounded models can zero-shot adapt to updated, unseen knowledge graphs. |
| Approach: | They propose a task to apply dynamic knowledge graphs to neural conversation models . they propose 'dyKgChat' that selects an output from two networks at each time step . |
| Outcome: | The proposed model outperforms existing knowledge-grounded conversation models in evaluation metrics. |
Open-Domain Conversational Question Answering with Historical Answers (2022.findings-aacl)
Copied to clipboard
| Challenge: | Existing approaches to conversational question answering are limited due to the large number of candidate documents. |
| Approach: | They propose a model that leverages historical answers to boost retrieval performance . they propose to use open-domain conversational question answering to solve these problems . |
| Outcome: | The proposed model outperforms baseline models in extractive and generative reader settings on OR-QuAC dataset. |