Papers by Buru Chang
The RAG Paradox: A Black-Box Attack Exploiting Unintentional Vulnerabilities in Retrieval-Augmented Generation Systems (2025.findings-emnlp)
Copied to clipboard
| Challenge: | Existing approaches to degrade retrieval-augmented generation systems rely on unrealistic assumptions in which external attackers have access to internal components such as the retriever. |
| Approach: | They propose a black-box attack based on the RAG paradox, a structural vulnerability that emerges from the system’s effort to enhance trust by revealing both the retrieved documents and their sources to users. |
| Outcome: | The proposed method significantly degrades system performance without internal access while generating natural-looking poisoned documents. |
SHARE: Shared Memory-Aware Open-Domain Long-Term Dialogue Dataset Constructed from Movie Script (2025.acl-long)
Copied to clipboard
| Challenge: | Antoine de Saint-Exupéry Memory in dialogue plays a crucial role in building relationships and facilitating the ongoing conversation. |
| Approach: | They propose a long-term dialogue dataset named SHARE that includes shared memories between two individuals. |
| Outcome: | The proposed dataset makes long-term dialogues more engaging and sustainable . it includes summaries of persona information and events of two individuals . |
Meet Your Favorite Character: Open-domain Chatbot Mimicking Fictional Characters with only a Few Utterances (2022.naacl-main)
Copied to clipboard
| Challenge: | In this paper, we consider mimicking fictional characters as a promising direction for building engaging conversation models. |
| Approach: | They propose a task where only a few utterances of each fictional character are available to generate responses mimicking them. |
| Outcome: | The proposed method generates responses better reflecting the style of fictional characters than baseline methods. |
HarDBench: A Benchmark for Draft-Based Co-Authoring Jailbreak Attacks for Safe Human–LLM Collaborative Writing (2026.acl-long)
Copied to clipboard
| Challenge: | Large language models are increasingly used as coauthors in collaborative writing . however, this capability poses a serious safety risk . |
| Approach: | They propose a safety-utility balanced alignment approach to train LLMs to refuse harmful completions while remaining helpful on benign drafts. |
| Outcome: | The proposed method reduces harmful outputs without degrading performance on co-authoring capabilities. |
“Killing Me” Is Not a Spoiler: Spoiler Detection Model using Graph Neural Networks with Dependency Relation-Aware Attention Mechanism (2021.eacl-main)
Copied to clipboard
| Challenge: | Several attention-based spoiler detection models are insufficient for utilizing dependency relations between context words. |
| Approach: | They propose a new spoiler detection model called SDGNN that uses syntax-aware graph neural networks to detect dependency relations between context words. |
| Outcome: | The proposed model outperforms existing models on two real-world benchmark datasets. |
Measuring and Improving Semantic Diversity of Dialogue Generation (2022.findings-emnlp)
Copied to clipboard
| Challenge: | Existing evaluation metrics for response diversity do not capture the semantic diversity of generated responses. |
| Approach: | They propose an automatic evaluation metric to measure the semantic diversity of generated responses . they show that it captures human judgments better than existing diversity metrics . |
| Outcome: | The proposed metric captures human judgments on response diversity better than existing lexical diversity metrics. |
Alignment Data Map for Efficient Preference Data Selection and Diagnosis (2026.findings-acl)
Copied to clipboard
| Challenge: | constructing high-quality preference datasets faces scalability challenges due to prohibitive cost and complexity of human annotation. |
| Approach: | They propose a tool to identify and select effective preference data by LLM-as-a-judge, explicit reward model, and reference-based approaches. |
| Outcome: | The proposed tool reduces annotation costs while preserving alignment effectiveness. |
Distilling the Knowledge of Large-scale Generative Models into Retrieval Models for Efficient Open-domain Conversation (2021.findings-emnlp)
Copied to clipboard
| Challenge: | generative models are less practical for building real-time conversation systems due to high latency and large memory footprint. |
| Approach: | They propose a method that preserves the efficiency of a retrieval model while leveraging the conversational ability of generative models. |
| Outcome: | The proposed method preserves the efficiency of a retrieval model while leveraging the conversational ability of generative models. |