Papers by Rui Qiu

17 papers
AgentGym: Evaluating and Training Large Language Model-based Agents across Diverse Environments (2025.acl-long)

Copied to clipboard

Challenge: Large language models (LLMs) are promising foundations to build generally-capable agents . however, the community lacks a unified interactive framework that covers diverse environments for comprehensive evaluation of agents.
Approach: They propose a framework that features 7 real-world scenarios, 14 environments, and 89 tasks for unified, real-time, and concurrent agent interaction.
Outcome: The proposed framework features 7 real-world scenarios, 14 environments, and 89 tasks for unified, real-time, and concurrent agent interaction.
Multi-Programming Language Sandbox for LLMs (2025.acl-demo)

Copied to clipboard

Challenge: MPLSandbox is an out-of-the-box multi-programming language sandbox designed to provide unified and comprehensive feedback from compiler and analysis tools for Large Language Models (LLMs).
Approach: They propose a multi-programming language sandbox that provides unified feedback from compilers and analysis tools for Large Language Models.
Outcome: The proposed multi-language sandbox can provide comprehensive feedback from compilers and analysis tools for large language models (LLMs).
Beyond Surface-Level Detection: Towards Cognitive-Driven Defense Against Jailbreak Attacks via Meta-Operations Reasoning (2026.acl-long)

Copied to clipboard

Challenge: Existing defenses rely on shallow pattern matching, which struggles to generalize to novel and unseen attack strategies.
Approach: They propose a framework which emulates human cognitive reasoning through a structured reasoning chain.
Outcome: The proposed framework achieves state-of-the-art performance and exhibits strong generalization to unseen attacks.
BaitAttack: Alleviating Intention Shift in Jailbreak Attacks via Adaptive Bait Crafting (2024.emnlp-main)

Copied to clipboard

Challenge: Existing attacks focus on meticulously constructing prompts to disguise harmful intentions . however, incorporation of disguising prompts may incur the challenge of "intention shift"
Approach: They propose a jailbreak attack component, BaitAttack, to alleviate the effects of intention shift . Bait provides a response to the query, prompting LLMs to rectify or supplement the knowledge within the bait .
Outcome: The proposed component, BaitAttack, reduces the effects of intention shift within jailbreak attacks.
ProLongVid: A Simple but Strong Baseline for Long-context Video Instruction Tuning (2025.emnlp-main)

Copied to clipboard

Challenge: Existing approaches to adapt image-focused models for video understanding have not been successful in analyzing long video sequences.
Approach: They propose a video instruction dataset that outperforms existing video instruction data for fine-tuning MLLMs by incrementally increasing input context length.
Outcome: The proposed model outperforms existing models on video benchmarks and outperformed proprietary models on VideoMME even with a compact 7B model.
Demons in the Detail: On Implementing Load Balancing Loss for Training Specialized Mixture-of-Expert Models (2025.acl-long)

Copied to clipboard

Challenge: Existing Mixture-of-Experts training frameworks use a micro-batch to calculate LBL . micro-batches are restricted to a single sequence, preventing expert specialization .
Approach: They propose to use a global-batch to loosen the load balance constraint for MoEs models . they propose to synchronize fi across micro-batches and then use it to calculate the LBL .
Outcome: The proposed global-batch LBL improves the domain specialization of experts . the micro-battery LBL is almost at the sequence level, and the router is pushed to distribute the token evenly .
Multimedia Event Extraction with LLM Knowledge Editing (2025.emnlp-main)

Copied to clipboard

Challenge: Existing multimodal event extraction methods focus on weakly aligning features from wellpretrained unimodal encoders, resulting in redundant feature perception.
Approach: They propose a multimodal event extraction strategy with a redundant feature selection mechanism that enhances event understanding ability of multimodal large language models.
Outcome: The proposed method outperforms the state-of-the-art (SOTA) baselines on the M2E2 benchmark.
Arithmetic Control of LLMs for Diverse User Preferences: Directional Preference Alignment with Multi-Objective Rewards (2024.acl-long)

Copied to clipboard

Challenge: Reinforcement Learning from Human Feedback (RLHF) relies on scalar rewards to capture user preferences.
Approach: They propose a framework that integrates multi-objective reward modeling to represent diverse preference profiles.
Outcome: The proposed method improves performance across reward objectives and targets.
Completing A Systematic Review in Hours instead of Months with Interactive AI Agents (2025.acl-long)

Copied to clipboard

Challenge: Systematic reviews (SRs) are vital for evidence-based practice in high stakes disciplines, such as healthcare.
Approach: They propose a human-centered interactive AI agent powered by large language models that partitions a large literature corpus based on semantics and employs .
Outcome: InsightAgent improves quality of synthesized SRs by 27.2%, reaching 79.7% of human-written quality.
Beyond Modality Collapse: Taming Guided Modality Entropy for Omni-modal Emotion Reasoning (2026.findings-acl)

Copied to clipboard

Challenge: EmoOmni is a data paradigm for omni-modal large language models that can be used for emotion reasoning.
Approach: They propose a data paradigm that interleaves guided tokens into reasoning traces to enforce structured evidence extraction.
Outcome: The proposed paradigm over-relys on a dominant modality while neglecting complementary cues.
VAPO: End-to-end Slide-Enhanced Speech Recognition with Omni-modal Large Language Models (2026.acl-long)

Copied to clipboard

Challenge: Current Automatic Speech Recognition models, such as Whisper, have demonstrated impressive performance in general domains, but their accuracy often deteriorates significantly in specialized scenarios.
Approach: They propose a visually-anchored policy optimization approach to decouple visual perception from auditory processing to optimize the model's inference process.
Outcome: The proposed model eliminates visual interference and achieves state-of-the-art performance on SlideASR-Bench and public datasets.
Are Training Samples Correlated? Learning to Generate Dialogue Responses with Multiple References (P19-1)

Copied to clipboard

Challenge: Existing approaches to open-domain dialogue generation ignore the nature of 1-to-1 mapping that there may exist multiple valid responses corresponding to the same query.
Approach: They propose to model open-domain dialogue generation using 1-to-1 mapping . they first extract common features of different responses and then combine them with distinctive features to generate multiple diverse and appropriate responses.
Outcome: The proposed model outperforms existing models on automatic and human evaluations.
ABC-Bench: Benchmarking Agentic Backend Coding in Real-World Development (2026.findings-acl)

Copied to clipboard

Challenge: Large Language Models (LLMs) have redefined the role of AI in software engineering . current benchmarks focus on localized code generation, but neglect dynamic, full-process requirements of real-world engineering.
Approach: They propose a benchmark to evaluate agentic backend coding within a realistic, executable workflow.
Outcome: The ABC-Bench benchmark evaluates agentic backend coding within a realistic, executable workflow.
ACSE: An Ancient Character Semantic-Aware Embedding for Large Language Models (2026.findings-acl)

Copied to clipboard

Challenge: Existing studies on pre-Qin documents are insufficient to understand ancient characters . ancient characters have a low level of digitization and training corpora are extremely scarce .
Approach: They propose a semantic-aware embedding for ancient Chinese characters that integrates glyphs and lexicality into modern Chinese semantic space.
Outcome: The proposed model integrates glyph and lexicality of ancient characters and maps them to the modern Chinese semantic space.
Do Not Guess, Verify: Logic-Guided Adaptive Reasoning for Multimodal Misinformation Detection (2026.findings-acl)

Copied to clipboard

Challenge: Existing multimodal misinformation detection paradigms rely on passive aggregation of multimodal features and social signals.
Approach: They propose a verification-oriented framework that integrates large vision–language models into multimodal misinformation detection through explicit rationale-guided reasoning.
Outcome: The proposed framework outperforms state-of-the-art methods on multimodal misinformation detection benchmarks while significantly reducing computational cost.
Investigating and Enhancing Vision-Audio Capability in Omnimodal Large Language Models (2025.findings-acl)

Copied to clipboard

Challenge: Recent years have witnessed significant advancements in large language models (LLMs) but still struggle with integrating vision and audio.
Approach: They propose a self-knowledge distillation method to improve vision-audio capabilities of OLLMs by learning from the vision-text components.
Outcome: The proposed method improves vision-audio capabilities of OLLMs by learning from vision-text components, which improves interaction between audio and images and results in improved performance on multimodal tasks.
TextFlint: Unified Multilingual Robustness Evaluation Toolkit for Natural Language Processing (2021.acl-demo)

Copied to clipboard

Challenge: Existing approaches to textual robustness evaluation focus on slightly modifying the input data, which maintains the original meaning and results in a different prediction.
Approach: They propose a multilingual robustness evaluation toolkit for NLP that integrates universal text transformations, task-specific transformations and adversarial attack.
Outcome: The toolkit includes universal text transformation, task-specific transformation, adversarial attack, subpopulation, and their combinations to provide comprehensive robustness analyses.

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