Papers by Caren Han
CONDA: a CONtextual Dual-Annotated dataset for in-game toxicity understanding and detection (2021.findings-acl)
Copied to clipboard
Henry Weld, Guanghao Huang, Jean Lee, Tongshu Zhang, Kunze Wang, Xinghong Guo, Siqu Long, Josiah Poon, Caren Han
| Challenge: | Existing toxic language detection models focus on the single utterance level without deeper understanding of context. |
| Approach: | They propose a dataset for in-game toxic language detection enabling joint intent classification and slot filling analysis, which is the core task of Natural Language Understanding (NLU). |
| Outcome: | The proposed framework handles utterance and token-level patterns, and rich contextual chatting history. |
ChuLo: Chunk-Level Key Information Representation for Long Document Understanding (2025.findings-acl)
Copied to clipboard
| Challenge: | Traditional approaches to truncate inputs, sparse self-attention, and chunking often lead to information loss and hinder the model’s ability to capture long-range dependencies. |
| Approach: | They propose a novel chunk representation method that uses unsupervised keyphrase extraction to group input tokens to retain core document content while reducing input length. |
| Outcome: | The proposed method minimizes information loss and improves the efficiency of Transformer-based models. |
SCO-VIST: Social Interaction Commonsense Knowledge-based Visual Storytelling (2024.eacl-long)
Copied to clipboard
| Challenge: | Visual storytelling aims to automatically generate a coherent story based on a given image sequence. |
| Approach: | They propose a framework that represents the image sequence as a graph with objects and relations that includes human action motivation and its social interaction commonsense knowledge. |
| Outcome: | The proposed framework produces stories superior across multiple metrics in terms of visual grounding, coherence, diversity, and humanness, per both automatic and human evaluations. |
MAGIC-VQA: Multimodal And Grounded Inference with Commonsense Knowledge for Visual Question Answering (2025.findings-acl)
Copied to clipboard
| Challenge: | Existing Large Vision-Language Models (LVLMs) lack integrated commonsense knowledge . lack of integrated common knowledge limits their robustness and accuracy in VQA . |
| Approach: | They propose a framework to enhance multimodal inference by integrating commonsense reasoning. |
| Outcome: | MAGIC-VQA improves comprehensive benchmark datasets, surpassing existing models in tasks requiring advanced commonsense reasoning. |
Re-Temp: Relation-Aware Temporal Representation Learning for Temporal Knowledge Graph Completion (2023.findings-emnlp)
Copied to clipboard
| Challenge: | Existing models ignore ability to skip irrelevant snapshots according to entity-related relations in query . TKGC is difficult and even large-scale pre-trained language models such as gist ignore explicit temporal information. |
| Approach: | They propose a model that leverages explicit temporal embedding as input to skip unnecessary information for prediction. |
| Outcome: | The proposed model outperforms all state-of-the-art models on six datasets . it incorporates skip information flow after each timestamp to skip unnecessary information . |
Detect All Abuse! Toward Universal Abusive Language Detection Models (2020.coling-main)
Copied to clipboard
| Challenge: | Existing work on online abusive language detection focused on detecting a single abusive language problem in a domain, like Twitter, but none of them was successfully transferable to general ALD in different online communities. |
| Approach: | They propose a generic ALD framework that can address multiple types of ALD tasks across different domains and use a textual graph embedding to analyse the user’s linguistic behaviour. |
| Outcome: | The proposed framework surpasses the current state-of-the-art ALD algorithms across seven datasets covering multiple aspects of abusive language and different online community domains. |
3MVRD: Multimodal Multi-task Multi-teacher Visually-Rich Form Document Understanding (2024.findings-acl)
Copied to clipboard
| Challenge: | Existing models for visually rich document understanding do not account for the diverse carriers of document versions and their associated noises. |
| Approach: | They propose a multimodal, multi-task, multiteacher joint-grained knowledge distillation model for visually-rich form document understanding. |
| Outcome: | The proposed model outperforms baselines on a comprehensive evaluation of public datasets showing it can handle complex structures and content of visually-rich forms. |
‘No’ Matters: Out-of-Distribution Detection in Multimodality Multi-Turn Interactive Dialogue Download PDF (2025.findings-acl)
Copied to clipboard
| Challenge: | Out-of-distribution (OOD) detection is essential for multimodal learning systems . a novel scoring framework is proposed to efficiently detect OOD in multi-round long dialogues . |
| Approach: | They propose a scoring framework that integrates visual language models with a score framework that detects OOD in two key scenarios. |
| Outcome: | The proposed framework detects OOD in two key scenarios: mismatches between dialogue and image input pair and previously unseen labels. |
RoViST: Learning Robust Metrics for Visual Storytelling (2022.findings-naacl)
Copied to clipboard
| Challenge: | Visual storytelling is the task of generating a story paragraph that describes a given image sequence. |
| Approach: | They propose 3 evaluation metrics sets that analyze which aspects we would look for in a good story . they compare their correlation with human judgement scores on a sample of machine stories . |
| Outcome: | The proposed evaluation metrics outperform other metrics on human correlation on a sample of machine stories from state-of-the-art models. |
EVOTOOL: Self-Evolving Tool-Use Policy Optimization in LLM Agents via Blame-Aware Mutation and Diversity-Aware Selection (2026.acl-long)
Copied to clipboard
| Challenge: | Existing approaches to optimize tool-use policies are monolithic and prone to entangling behaviors. |
| Approach: | They propose a framework that decomposes agent’stool-use policy into four modules and improves them via three mechanisms. |
| Outcome: | The proposed framework outperforms strong baselines on bothGPT-4.1 and Qwen3-8B while maintaining superior efficiency and transferability. |
MIDAS: Multi-level Intent, Domain, And Slot Knowledge Distillation for Multi-turn NLU (2025.findings-naacl)
Copied to clipboard
| Challenge: | Existing Large Language Models (LLMs) can generate coherent text, but they struggle to recognise user intent behind queries. |
| Approach: | They propose a novel approach leveraging multi-level intent, domain, and slot knowledge distillation for multi-turn NLU. |
| Outcome: | The proposed model improves multi-turn conversation understanding by integrating teacher teachers into a student model. |
VICTR: Visual Information Captured Text Representation for Text-to-Vision Multimodal Tasks (2020.coling-main)
Copied to clipboard
| Challenge: | Existing text-to-image generation models focus on generating high resolution images and neglect understanding text descriptions. |
| Approach: | They propose a visual contextual text representation which captures rich visual semantic information of objects from text input. |
| Outcome: | The proposed visual contextual text representation improves on the state-of-the-art models. |
A Training-Free Length Extrapolation Approach for LLMs: Greedy Attention Logit Interpolation (2025.emnlp-main)
Copied to clipboard
| Challenge: | Existing solutions to problem of positional out-of-distribution (O.O.D.) are inefficient, redundant, and lack local positional information. |
| Approach: | They propose a training-free method that greedily reuses pretrained positional intervals and interpolates attention logits to eliminate outliers. |
| Outcome: | The proposed method achieves stable and superior performance across long-context tasks without requiring input-length-specific tuning. |