Papers by Dong Yan

38 papers
CART: A Generative Cross-Modal Retrieval Framework With Coarse-To-Fine Semantic Modeling (2025.acl-long)

Copied to clipboard

Challenge: Cross-modal retrieval tasks are used to retrieve data from one modality or another based on a query from another modality.
Approach: They propose a generative cross-modal retrieval framework based on coarse-to-fine semantic modeling . they propose combining K-Means and RQ-VAE to discretize multimodal data into token sequences that support autoregressive generation.
Outcome: The proposed framework achieves excellent performance and efficiency in multimodal retrieval tasks.
SILC-EFSA: Self-aware In-context Learning Correction for Entity-level Financial Sentiment Analysis (2025.coling-main)

Copied to clipboard

Challenge: Currently, most sentiment analysis corpora use sequence-level annotation.
Approach: They propose a two-stage approach to financial entity-level sentiment analysis called Self-aware In-context Learning Correction.
Outcome: The proposed approach achieves state-of-the-art on the largest English and Chinese financial entity-level sentiment analysis datasets to date.
Multiplex Word Embeddings for Selectional Preference Acquisition (D19-1)

Copied to clipboard

Challenge: Existing word embeddings are limited in their ability to represent fixed vectors . instead, they incorporate relational dependencies of different words into their embeddables - a limitation that is addressed by a multiplex model .
Approach: They propose a word embedding model which incorporates relational dependencies of different words into their embeddables.
Outcome: The proposed model can be easily extended according to various relations among words.
MMedAgent: Learning to Use Medical Tools with Multi-modal Agent (2024.findings-emnlp)

Copied to clipboard

Challenge: Multi-modal Large Language Models (MLLMs) exhibit limited generality and often fall short when compared to specialized models.
Approach: They propose a multi-modal medical agent that picks the most suitable medical tools based on user inputs.
Outcome: The proposed agent performs better than open-source models and the closed-source model, GPT-4o.
Topology Matters: Measuring Memory Leakage in Multi-Agent LLMs (2026.findings-acl)

Copied to clipboard

Challenge: Multi-agent LLMs are rapidly moving from prototype to real-world use . network topology is a first-order security parameter in multi-aggent systems .
Approach: They propose a framework for comparing topology-conditioned memory leakage in multi-agent LLM systems.
Outcome: The proposed framework evaluates topology-conditioned memory leakage in multi-agent LLM systems.
AnyGPT: Unified Multimodal LLM with Discrete Sequence Modeling (2024.acl-long)

Copied to clipboard

Challenge: Existing language models that use discrete representations for unified processing of various modalities are limited to text generation and do not include multimodal output.
Approach: They propose a multimodal language model that utilizes discrete representations for unified processing of various modalities.
Outcome: The proposed model can be trained stably without any alterations to existing models or training paradigms.
DeepGuard: Secure Code Generation via Multi-Layer Semantic Aggregation (2026.acl-long)

Copied to clipboard

Challenge: Large Language Models (LLMs) can replicate insecure patterns from training data.
Approach: They propose a framework that leverages distributed security-relevant cues by aggregating representations from multiple upper layers via an attention-based module.
Outcome: Experiments show that the framework improves the secure-and-correct generation rate by 11.9% over baselines.
“I’ve Decided to Leak”: Probing Internals Behind Prompt Leakage Intents (2025.emnlp-main)

Copied to clipboard

Challenge: Large language models (LLMs) exhibit prompt leakage vulnerabilities, raising intellectual property and confidentiality concerns.
Approach: They use probing techniques to capture LLMs’ intent-related internal representations and show that they internalize prompt leakage intents in their hidden states before generating tokens.
Outcome: The proposed probes achieve 90%+ AUROC across all tested models, even when applied to new system prompts and attacks.
Large Language Models as Zero-shot Dialogue State Tracker through Function Calling (2024.acl-long)

Copied to clipboard

Challenge: Large language models (LLMs) are increasingly prevalent in conversational systems due to their advanced understanding and generative capabilities in general contexts.
Approach: They propose a method for solving dialogue state tracking (DST) with large language models through function calling.
Outcome: The proposed approach improves zero-shot DST, allowing adaptation to diverse domains without extensive data collection or model tuning.
Covariance Matrix-Driven Image Channel Allocation for Multimodal Fake News Detection (2026.findings-acl)

Copied to clipboard

Challenge: Existing methods focus on designing efficient multimodal fusion frameworks to bridge the semantic gap between images and texts.
Approach: They propose a covariance matrix-driven image channel allocation method that expands the number of original channel maps and assigns importance scores to the expanded channel maps.
Outcome: The proposed method achieves state-of-the-art on three public multimodal fake news detection benchmark datasets.
SocialEval: Evaluating Social Intelligence of Large Language Models (2025.acl-long)

Copied to clipboard

Challenge: Existing work on LLMs does not address their social intelligence (SI) and their discrepancy with humans.
Approach: They propose a script-based bilingual SI benchmark that integrates outcome-oriented goal achievement evaluation and process-oriented interpersonal ability evaluation by manually crafting narrative scripts.
Outcome: The proposed model is based on a script-based bilingual evaluation paradigm that integrates outcome- and process-oriented evaluation by manually crafting narrative scripts.
What If Consensus Lies? Selective-Complementary Reinforcement Learning at Test Time (2026.acl-long)

Copied to clipboard

Challenge: Existing TTRL methods rely on positive pseudo-labeling strategies to enhance reasoning capabilities.
Approach: They propose a test-time reinforcement learning framework that mitigates label noise amplification by deriving pseudo-rewards from majority voting consensus.
Outcome: The proposed framework mitigates label noise amplification by implementing selective positive pseudo-labeling and entropy-gated negative p-labeled pruning.
MatPlotAgent: Method and Evaluation for LLM-Based Agentic Scientific Data Visualization (2024.findings-acl)

Copied to clipboard

Challenge: Scientific data visualization is an essential process in research, but its use of large language models remains unexplored.
Approach: They propose a model-agnostic LLM agent framework to automate scientific data visualization tasks.
Outcome: The proposed framework improves performance of commercial and open-source models.
Representation Interventions Enable Lifelong Knowledge Memory Control in LLMs (2026.acl-long)

Copied to clipboard

Challenge: Large language models (LLMs) produce outdated or inaccurate content. Updating their knowledge efficiently and accurately without costly retraining is a major challenge.
Approach: They propose a robust and scalable method that treats knowledge control as interventions within the model’s representation space.
Outcome: The proposed method achieves fine-grained control over complex, unstructured knowledge while maintaining general utility with frozen base weights.
Reward Modeling Requires Automatic Adjustment Based on Data Quality (2024.findings-emnlp)

Copied to clipboard

Challenge: Reinforcement Learning from Human Feedback (RLHF) is a method for aligning language models with human values.
Approach: They propose a method that automatically adjusts reward modeling based on data quality . they use preference data to train a reward model that is more aligned with human values .
Outcome: The proposed method stabilizes reward model training and significantly improves alignment performance on human preference datasets.
DeepGen: Diverse Search Ad Generation and Real-Time Customization (2022.emnlp-demos)

Copied to clipboard

Challenge: Existing systems that generate ads manually are not effective in generating ad copy and generating millions of ads for large businesses.
Approach: They propose a system that generates fluent ads from advertiser’s web pages in an abstractive fashion and solves practical issues such as factuality and inference speed.
Outcome: The proposed system generates fluent ads from advertiser’s web pages in an abstractive fashion and solves practical issues such as factuality and inference speed.
Entity-level Interaction via Heterogeneous Graph for Multimodal Named Entity Recognition (2022.findings-emnlp)

Copied to clipboard

Challenge: Existing methods for name-based entity recognition neglect the integrity of entity semantics and conduct cross-modal interaction at token-level.
Approach: They propose a multimodal named entity recognition model that captures visual information and fuses it into tokens to rid non-entity tokens of visual noise.
Outcome: The proposed model captures entity-related visual information and fuses it into tokens . it eliminates visual noise and makes non-entity tokens easily misidentified as entities .
Multi-Scale Spectral Selection and Entropy-Guided Uncertainty Fusion for Multimodal Rumor Detection (2026.findings-acl)

Copied to clipboard

Challenge: Existing methods for multimodal content detection fail to capture cross-modal semantic inconsistencies and ignore inherent noise in multimodal features.
Approach: They propose a multimodal rumor detection method based on a frequency domain spectral selection method and entropy-guided uncertainty fusion method to capture cross-modal semantic inconsistencies.
Outcome: The proposed method outperforms state-of-the-art methods in multimodal rumor detection . it shows stronger detection capability and robustness on multiple datasets .
Persona-E²: A Human-Grounded Dataset for Personality-Shaped Emotional Responses to Textual Events (2026.acl-long)

Copied to clipboard

Challenge: A critical bottleneck is the lack of ground-truth human data to link personality traits to emotional shifts.
Approach: They propose a large-scale dataset to capture reader-based emotional variations across news, social media, and life narratives.
Outcome: The proposed model captures reader-based emotional variations across news, social media, and life narratives.
HiTab: A Hierarchical Table Dataset for Question Answering and Natural Language Generation (2022.acl-long)

Copied to clipboard

Challenge: Existing studies on table reasoning focus on flat tables and hierarchical tables . a new dataset, HiTab, aims to examine numerical reasoning over hierarchic tables based on hierarchically structured tables - a strong challenge for existing baselines and a valuable benchmark for future research.
Approach: They propose a hierarchical question answering and natural language generation dataset to study hierarchic tables.
Outcome: The proposed model shows that it is effective in QA and natural language generation over hierarchical tables.
TASO: Task-Aligned Sparse Optimization for Parameter-Efficient Model Adaptation (2025.emnlp-main)

Copied to clipboard

Challenge: Existing studies have shown that LoRA introduces substantial parameter redundancy, which not only increases the number of trainable parameters but also hinders the effectiveness of fine-tuning.
Approach: They propose a method that leverages importance information from the pretrained model’s weights to mitigate LoRA redundancy.
Outcome: The proposed method significantly reduces the number of trainable parameters required for task adaptation while providing a task-aligned perspective for LoRA redundancy reduction.
Enhancing Multimodal Unified Representations for Cross Modal Generalization (2025.findings-acl)

Copied to clipboard

Challenge: Existing studies on discrete unified representations overlook important distinctions between different dimensions of features.
Approach: They propose to use a codebook to optimize unified representations from pretraining and fine- and coarse-grained disentangling to optimize the representations.
Outcome: The proposed methods improve the interpretability of multimodal unified representations . they use training-free optimization of codebook and fine and coarse cross-modal disentangling .
Vision-Language Introspection: Mitigating Overconfident Hallucinations in MLLMs via Interpretable Bi-Causal Steering (2026.acl-long)

Copied to clipboard

Challenge: Existing approaches to overcome object hallucination are limited . Existing mitigations include costly retraining and a training-free inference framework .
Approach: They propose a training-free inference framework that simulates a metacognitive self-correction process.
Outcome: The proposed framework reduces object hallucination rates by 12.67% on MMHal-Bench and improves accuracy by 5.8% on POPE.
XL-NBT: A Cross-lingual Neural Belief Tracking Framework (D18-1)

Copied to clipboard

Challenge: a multi-lingual approach to training dialog systems is expensive and tedious, but it can be useful for cross-lingual support.
Approach: They propose to annotate data for multiple languages and train a multi-lingual dialog system for each language.
Outcome: The proposed framework bypasses the expensive human annotation and achieves promising results.
Attention Entropy is a Key Factor: An Analysis of Parallel Context Encoding with Full-attention-based Pre-trained Language Models (2025.acl-long)

Copied to clipboard

Challenge: Large language models have demonstrated remarkable performance across a wide range of language tasks due to their remarkable ability in context modeling.
Approach: They propose to use parallel context encoding to reduce attention entropy by incorporating attention sinks and selective mechanisms to reduce irregular attention . they also propose to incorporate attention sink mechanisms into the parallel encoded context to reduce the irregular attention.
Outcome: The proposed methods lower irregular attention entropy and narrow performance gaps.
Perovskite-LLM: Knowledge-Enhanced Large Language Models for Perovskite Solar Cell Research (2025.findings-emnlp)

Copied to clipboard

Challenge: a rapid advancement of perovskite solar cells has led to an exponential growth in research publications.
Approach: They propose a knowledge-enhanced system for perovskite solar cells that integrates three key components.
Outcome: The proposed system outperforms existing models in domain-specific knowledge retrieval and scientific reasoning tasks.
Finch: Benchmarking Finance & Accounting across Spreadsheet-Centric Enterprise Workflows (2026.findings-acl)

Copied to clipboard

Challenge: FinWorkBench evaluates real-world enterprise-grade finance and accounting workflows . a human evaluation of GPT 5.1 Pro passes only 38.4% of workflows, a study finds .
Approach: They propose a workflow construction process that combines LLM-assisted mining and expert annotation to build 172 composite workflows.
Outcome: The proposed process combines expert annotation with LLM-assisted mining of workflows from authentic enterprise environments.
Revisiting the Reliability of Language Models in Instruction-Following (2026.acl-long)

Copied to clipboard

Challenge: Several benchmarks have been proposed to measure instruction-following accuracy, but these scores do not translate to reliable services in real-world use.
Approach: They propose a new metric reliable@k and develop an automated pipeline to generate cousin prompts.
Outcome: The proposed model can be instantiated with cousin prompts and generates high-quality cousin prompt data.
AlignMMBench: Evaluating Chinese Multimodal Alignment in Large Vision-Language Models (2025.acl-long)

Copied to clipboard

Challenge: Existing benchmarks focus on basic abilities using nonverbal methods, such as yes-no and multiple-choice questions.
Approach: They propose a benchmark that provides more nuanced evaluations of alignment capabilities for large Vision-Language Models (VLMs) they use a rule-calibrated evaluator that exceeds GPT-4's evaluation ability and a “alignment score” to assess the robustness and stability of models across diverse prompts.
Outcome: The proposed benchmark covers 13 tasks across three categories and includes both single-turn and multi-turn dialogue scenarios.
RecBase: Generative Foundation Model Pretraining for Zero-Shot Recommendation (2025.emnlp-main)

Copied to clipboard

Challenge: Existing methods for addressing item-level user interests are lacking in cross-domain generalization . RecBase model is domain-agnostic and can be used to enhance recommender systems' effectiveness .
Approach: They propose a domain-agnostic foundational model pretrained with a recommendation-oriented objective that leverages a large-scale, heterogeneous, cross-domain corpus with unified textual representations and feature mappings to enhance cross- domain generalization.
Outcome: The proposed model matches or surpasses baselines in zero-shot and cross-domain recommendation tasks on eight real-world datasets.
ProTOD: Proactive Task-oriented Dialogue System Based on Large Language Model (2025.coling-main)

Copied to clipboard

Challenge: Existing task-oriented dialogue systems engage with users in a reactive manner, relying on a basic single-query mechanism and employing passive policy planning.
Approach: They propose a novel LLM-based proactive TOD framework to improve system proactivity and goal completion.
Outcome: The proposed framework improves system proactivity and goal completion rates by 10% while enhancing proactive engagement.
Deep-Reporter: Deep Research for Grounded Multimodal Long-Form Generation (2026.acl-long)

Copied to clipboard

Challenge: Recent agentic search frameworks are text-centric, overlooking multimodal evidence . a pressing task is multimodal long-form generation, a new paper argues .
Approach: They propose a unified agentic framework for grounded multimodal long-form generation.
Outcome: The proposed framework is based on a unified agentic framework for grounded multimodal long-form generation.
AdaTag: Multi-Attribute Value Extraction from Product Profiles with Adaptive Decoding (2021.acl-long)

Copied to clipboard

Challenge: Existing approaches to extract product attribute values are limited by knowledge sharing across different attributes.
Approach: They propose to use adaptive decoding to handle extraction of product attribute values by parameterizing the decoder with pretrained attribute embeddings, through a hypernetwork and a Mixture-of-Experts module.
Outcome: The proposed model is able to handle multiple attributes without sharing the entire network parameters across all attributes.
Knowledge-aware Pronoun Coreference Resolution (P19-1)

Copied to clipboard

Challenge: Existing models for pronoun coreference resolution only use triplets, the most common format for knowledge graphs.
Approach: They propose a model that leverages different types of knowledge to resolve pronoun coreference with a neural model.
Outcome: The proposed model outperforms state-of-the-art baselines on two datasets from different domains.
Cross-lingual Knowledge Graph Alignment via Graph Matching Neural Network (P19-1)

Copied to clipboard

Challenge: Existing approaches to cross-lingual knowledge graph (KG) alignment rely on entity embeddings derived from monolingual KG structural information.
Approach: They propose a topic entity graph to represent entities with contextual information in KGs.
Outcome: The proposed model outperforms state-of-the-art methods by a large margin.
Inverting the Shield: Systematically Generating Safety Tests from Policy Specifications (2026.acl-long)

Copied to clipboard

Challenge: Existing safety evaluation paradigms rely on constructed benchmarks or dynamic red-teaming to probe potential vulnerabilities.
Approach: They propose a framework that integrates specification-based software testing with AI safety.
Outcome: The proposed framework achieves higher coverage and attack success counts compared to baselines.
Reward Generalization in RLHF: A Topological Perspective (2025.findings-acl)

Copied to clipboard

Challenge: Existing alignment methods share a common topology of information flow, but their alternatives have not been thoroughly explored.
Approach: They propose a theory of reward generalization in reinforcement learning from human feedback . they propose induced Bayesian networks to model the impact of dataset topologies on reward generalisation .
Outcome: The proposed method achieves an average win rate of 65% on three NLP tasks.
NL2Formula: Generating Spreadsheet Formulas from Natural Language Queries (2024.findings-eacl)

Copied to clipboard

Challenge: Creating spreadsheet formulas remains a tedious and error-prone task for many end-users . a novel task is proposed to generate spreadsheet formulae from a user's NL query .
Approach: They propose a task to generate formulas that are grounded on a spreadsheet table given a Natural Language query as input.
Outcome: The proposed task generates formulas that are grounded on a spreadsheet table, given a natural language query as input.

What is GenGO?

GenGO is an NLP powered publication search system. It currenctly indexes 30k+ papers from ACL Anthology, and implements multi-aspect summarization, semantic search, and more!

Information

About
Limitations