Papers by Wenji Mao

18 papers
Dynamic Routing Transformer Network for Multimodal Sarcasm Detection (2023.acl-long)

Copied to clipboard

Challenge: Existing methods for multimodal sarcasm detection rely on fixed architectures to capture cross-modal incongruity.
Approach: They propose a method that uses dynamic paths to activate different routing transformer modules with hierarchical co-attention adapting to cross-modal incongruity.
Outcome: The proposed method is compared to state-of-the-art methods on a public dataset.
Cross-Lingual Cross-Target Stance Detection with Dual Knowledge Distillation Framework (2023.emnlp-main)

Copied to clipboard

Challenge: Existing studies on stance detection were conducted mainly in English due to the low-resource problem in most non-English languages.
Approach: They propose to use a cross-lingual teacher and a teacher to transfer knowledge from source to target language to bridge the discrepancy between languages.
Outcome: The proposed framework bridges the discrepancy between languages and generalizes the knowledge to unseen targets in target language.
ImaRA: An Imaginative Frame Augmented Method for Low-Resource Multimodal Metaphor Detection and Explanation (2025.findings-naacl)

Copied to clipboard

Challenge: Existing methods for multimodal metaphor detection neglect cross-domain and attribute similarity characteristics underlying multimodal understanding.
Approach: They propose an Imaginative FRame Augmented method for multimodal metaphor detection and explanation . they use a cross-modal imagination dataset rich in multimodal multimodal expressions .
Outcome: The proposed method outperforms existing methods with training data on two datasets.
Modeling Conceptual Attribute Likeness and Domain Inconsistency for Metaphor Detection (2023.emnlp-main)

Copied to clipboard

Challenge: Metaphor detection aims to distinguish between metaphorical and literal expressions in text.
Approach: They propose an attribute likeness and domain inconsistency learning framework for wordpair metaphor detection based on conceptual metaphor theory . they model attribute likeity with an attribute siamese network and devise a domain contrastive learning strategy to learn semantic inconsistentness of concepts in source and target domains .
Outcome: The proposed framework outperforms existing word-pair and token-level methods on four datasets.
One Unified Model for Diverse Tasks: Emotion Cause Analysis via Self-Promote Cognitive Structure Modeling (2025.naacl-long)

Copied to clipboard

Challenge: Existing models for emotion cause analysis overlook common ground rooted in cognitive emotion theories, in particular, the cognitive structure of emotions.
Approach: They propose a unified model capable of tackling diverse emotion cause analysis tasks . they propose 'self-promote mechanism' that constructs the emotion cognitive structure through LLM .
Outcome: The proposed model outperforms existing models and baselines on multiple emotion cause analysis tasks.
Effective Inter-Clause Modeling for End-to-End Emotion-Cause Pair Extraction (2020.acl-main)

Copied to clipboard

Challenge: Emotion-cause pair extraction aims to extract all emotion clauses coupled with their cause clauses from a given document.
Approach: They propose a one-step neural approach which emphasizes inter-clause modeling to perform end-to-end extraction.
Outcome: The proposed method outperforms existing methods in the extraction of emotion-cause pairs . it emphasizes inter-clause modeling to perform end-to-end extraction .
Generative Adversarial Training with Perturbed Token Detection for Model Robustness (2023.emnlp-main)

Copied to clipboard

Challenge: Existing adversarial training methods use discrete tokens to deceive models . current approaches use embeddings, whereas actual text-based training uses discrete text tokens.
Approach: They propose a framework that integrates gradient-based learning, adversarial example generation and perturbed token detection to enhance adversariarial robustness.
Outcome: The proposed framework surpasses the state-of-the-art results of ChatGPT by 10% in average accuracy.
PromISe: Releasing the Capabilities of LLMs with Prompt Introspective Search (2024.lrec-main)

Copied to clipboard

Challenge: Existing evaluation benchmarks for large language models use uniform manual prompts, resulting in underestimation of performance.
Approach: They propose a prompt introspective search framework that integrates self-introspect and self-refine to unlock the capabilities of LLMs.
Outcome: The proposed framework significantly boosts the performance of 12 well-known LLMs compared to baseline methods.
DEMO: Reframing Dialogue Interaction with Fine-grained Element Modeling (2025.findings-acl)

Copied to clipboard

Challenge: Large language models (LLMs) enabled dialogue systems are one of the central modes in human-machine interaction.
Approach: They propose a benchmark task for dialogue element MOdeling and Element Awareness and a new benchmark for dialogue agent interaction that allows the agent to model dialogue elements via imitation learning.
Outcome: The proposed agent performs well in both dialogue element modeling and out-of-domain tasks.
Perspective-driven Preference Optimization with Entropy Maximization for Diverse Argument Generation (2025.findings-emnlp)

Copied to clipboard

Challenge: Argument generation with diverse perspectives is essential for fostering balanced discourse and mitigating bias.
Approach: They propose a Perspective-aware Preference Optimization with Entropy Maximization framework for diverse argument generation.
Outcome: The proposed framework enhances perspective diversity through preference optimization based on the constructed preference dataset .
A Theory Guided Scaffolding Instruction Framework for LLM-Enabled Metaphor Reasoning (2024.naacl-long)

Copied to clipboard

Challenge: Existing methods for metaphor detection and reasoning struggle to explain the underlying reasoning process behind the metaphorical/literal judgment.
Approach: They propose a Theory guided Scaffolding Instruction framework that instructs an LLM to infer the underlying reasoning process of metaphor detection guided by metaphor theories for the first time.
Outcome: The proposed method significantly outperforms both the LLM-based reasoning methods and the SOTA methods in metaphor detection.
TARA: Token-level Attribute Relation Adaptation for Multi-Attribute Controllable Text Generation (2024.findings-emnlp)

Copied to clipboard

Challenge: Existing work on multi-attribute controllable text generation ignores interrelations of attributes . recent work defines attribute relations as promotive, but not fixed .
Approach: They propose a method that explicitly defines attribute relations as inhibtory for multi-attribute CTG . they propose 'tara' which employs token-level attribute relation adaptation and representation to generate text with the balanced multi-attribut .
Outcome: The proposed method generates text with the balanced multi-attribute control.
Target-Oriented Relation Alignment for Cross-Lingual Stance Detection (2023.findings-acl)

Copied to clipboard

Challenge: Existing work on cross-lingual stance detection has ignored the inconsistency in the occurrences and distributions of targets between languages, which consequently degrades the performance of stance detector in low-resource languages.
Approach: They propose a fine-grained method which considers both target-level associations and language-level alignments to learn the in-language and cross-language associations.
Outcome: The proposed method is compared with competing methods under variant settings and shows that it performs better in low-resource languages.
Bridging Word-Pair and Token-Level Metaphor Detection with Explainable Domain Mining (2024.acl-long)

Copied to clipboard

Challenge: Existing methods for word-pair metaphor detection provide intermediate explainable clues for detection results.
Approach: They propose a method to bridge word-pair and token-level metaphor detection by modeling word pairs as explainable intermediate information.
Outcome: The proposed method bridges word-pair and token-level metaphor detection by using word pairs . it provides intermediate explainable clues for the detection results, but this is a challenge .
An LLM-Enabled Knowledge Elicitation and Retrieval Framework for Zero-Shot Cross-Lingual Stance Identification (2024.findings-emnlp)

Copied to clipboard

Challenge: Existing research is conducted in monolingual setting on English datasets, whereas in other low-resource languages, it lacks sufficient data for training quality stance detection models.
Approach: They propose a knowledge elicitation and retrieval framework that leverages the capability of large language models for stance knowledge acquisition and matches the target language input to the most relevant stance information.
Outcome: The proposed framework improves on multilingual datasets and competitive baselines.
Modeling Conversation Structure and Temporal Dynamics for Jointly Predicting Rumor Stance and Veracity (D19-1)

Copied to clipboard

Challenge: Existing methods to verify rumors are needed to identify false rumors.
Approach: They propose a hierarchical multi-task learning framework for jointly predicting rumor stance and veracity on Twitter that exploits the temporal dynamics of stance evolution.
Outcome: The proposed framework outperforms previous methods on two benchmark datasets showing that it can predict rumor stance and veracity.
Reasoning with Multimodal Sarcastic Tweets via Modeling Cross-Modality Contrast and Semantic Association (2020.acl-main)

Copied to clipboard

Challenge: Existing methods for sarcasm detection rely on text data, but are insufficient to detect multimodal sarcasm.
Approach: They propose a method for modeling cross-modality contrast in the associated context by constructing the Decomposition and Relation Network.
Outcome: The proposed model can detect sarcasm in multimodal tweets using a dataset .
Breaking the Impasse: Dual-Scale Evolutionary Policy Training for Social Language Agents (2026.acl-long)

Copied to clipboard

Challenge: Reinforcement Learning with Verifiable Rewards (RLVR) is effective for closed-ended tasks, but it is not applicable to open-ended social language games.
Approach: They propose a method that uses a time-scaled evolutionary perception mechanism to detect impasse by quantifying dual-scale value baseline divergence alongside match entropy.
Outcome: Experiments on multiple social language games show that the proposed method outperforms baselines and avoids policy degeneration.

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