Papers by Yiwen Zhang

25 papers
TagSpeech: End-to-End Multi-Speaker ASR and Diarization with Fine-Grained Temporal Grounding (2026.acl-long)

Copied to clipboard

Challenge: a critical ambiguity persists regarding what constitutes "joint ASR and diarization" a unified framework for multi-speaker ASR is proposed, but it is not yet clear what constitute "diarization."
Approach: They propose a unified LLM-based framework that uses Temporal Anchor Grounding for joint multi-speaker ASR and diarization.
Outcome: The proposed framework improves on AMI and AliMeeting benchmarks on speaker-content alignment . the proposed framework achieves consistent improvements in Diarization Error Rate over strong baselines .
DLTKG: Denoising Logic-based Temporal Knowledge Graph Reasoning (2025.findings-emnlp)

Copied to clipboard

Challenge: Current approaches to temporal knowledge representation face limited generalization to unseen facts and insufficient interpretability of reasoning processes.
Approach: They propose a framework that uses a denoising diffusion process to complete reasoning tasks . they propose introducing a noise source and historical conditionguiding mechanism to improve interpretability .
Outcome: The proposed framework outperforms state-of-the-art methods on three benchmark datasets.
Pause or Fabricate? Training Language Models for Grounded Reasoning (2026.findings-acl)

Copied to clipboard

Challenge: Large language models implicitly fabricate information when inputs are incomplete, causing confidence but unreliable conclusions.
Approach: They propose a framework for grounded reasoning under incomplete information that decomposes reasoning into two stages . they propose stage-specific rewards to penalize hallucinations, enabling models to detect gaps, stop proactively, and resume reasoning after clarification.
Outcome: The proposed framework improves premise detection and task success by 30% . it also reduces average response length by over 20% .
Unveiling and Consulting Core Experts in Retrieval-Augmented MoE-based LLMs (2024.emnlp-main)

Copied to clipboard

Challenge: Existing research seeks to enhance RAG performance by retrieving higher-quality documents or designing RAG-specific LLMs, but internal mechanisms that contribute to RAG’s effectiveness remain underexplored.
Approach: They propose to examine the internal mechanisms within the popular Mixture-of-Expert (MoE)-based LLMs and examine their ability to improve RAG by examining expert activations.
Outcome: The proposed method significantly improved the ability of Large Language Models (LLMs) to solve knowledge-intensive tasks.
Counteracting the Matthew Effect in Self-Improvement of LVLMs through Head-Tail Re-balancing (2026.acl-long)

Copied to clipboard

Challenge: Large vision language models have impressive reasoning capabilities across complex multimodal tasks.
Approach: They propose to use distribution-reshaping and trajectory-rebalancing to improve visual reasoning capabilities.
Outcome: Experiments on Qwen2-VL-7B-Instruct and InternVL2.5-4B models show that their methods outperform baselines by 3.86 points.
Mitigating Tail Narrowing in LLM Self-Improvement via Socratic-Guided Sampling (2025.naacl-long)

Copied to clipboard

Challenge: Large language models (LLMs) generate solutions themselves and iteratively train on filtered, high-quality rationales, but performance reaches a ceiling after a few iterations.
Approach: They propose a strategy to improve the efficiency of sampling heavy-tailed data by using Socratic-style guidance signals to help LLMs reasoning with complex queries.
Outcome: The proposed approach is effective on difficult queries and on held-out tasks, while requiring human supervision.
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.
Investigating Transfer Learning in Multilingual Pre-trained Language Models through Chinese Natural Language Inference (2021.findings-acl)

Copied to clipboard

Challenge: Multilingual transformers have been shown to have remarkable transfer skills in zero-shot settings.
Approach: They investigate cross-lingual transfer abilities of XLM-R for Chinese and English natural language inference using a large scale Chinese dataset.
Outcome: The proposed model trains on Chinese and English natural language inference datasets.
Polynomial Expansion Rank Adaptation: Enhancing Low-Rank Fine-Tuning with High-Order Interactions (2026.findings-acl)

Copied to clipboard

Challenge: Low-rank adaptation (LoRA) is a widely used strategy for efficient fine-tuning of large language models, but its strictly linear structure limits expressive capacity.
Approach: They propose a method that introduces structured polynomial expansion directly into the low-rank factor space.
Outcome: The proposed method outperforms state-of-the-art methods across diverse benchmarks.
DDPrompt: Differential Diversity Prompting in Large Language Models (2024.acl-short)

Copied to clipboard

Challenge: Large Language Models (LLMs) have shown that their reasoning ability can be enhanced through approaches like Chain-of-Thought (CoT) prompting.
Approach: They propose a method that generates differentially diverse reasoning paths for different types of questions by voting on the optimal prompts.
Outcome: The proposed method improves LLMs' reasoning ability on complex reasoning tasks by learning from demonstrations while keeping their parameters frozen.
CLUE: A Chinese Language Understanding Evaluation Benchmark (2020.coling-main)

Copied to clipboard

Challenge: Existing language evaluation benchmarks for English are limited to English . lack of such benchmarks makes it difficult to replicate success in other languages .
Approach: They introduce a large-scale Chinese language understanding evaluation benchmark . the benchmark uses a set of current state-of-the-art pre-trained Chinese models .
Outcome: The first large-scale Chinese Language Understanding Evaluation (CLUE) benchmark is released . the benchmark evaluates models across a wide range of tasks on original Chinese text . existing language evaluation benchmarks are mostly limited to English .
LONGAGENT: Achieving Question Answering for 128k-Token-Long Documents through Multi-Agent Collaboration (2024.emnlp-main)

Copied to clipboard

Challenge: Large language models (LLMs) have been successful in understanding language and processing text, but their cost prohibits their practical applications.
Approach: They propose a multi-agent collaboration method that breaks down lengthy documents into smaller, more manageable chunks and organizes the member agents to read their assigned chunks.
Outcome: The proposed method achieves 16.42% and 1.63% accuracy gains over existing models on single-hop and multi-hop QA settings.
LitVISTA: A Benchmark for Narrative Orchestration in Literary Text (2026.acl-long)

Copied to clipboard

Challenge: Existing large language models focus on causal coherence, neglecting the complex story arcs and orchestration inherent in human narratives.
Approach: They propose a high-dimensional framework for narrative orchestration that unifies human and model perspectives while jointly characterizing narrative function and structure in a common space.
Outcome: The proposed framework unifies human and model perspectives while jointly characterizing narrative function and structure in a common space.
OpenResearcher: Unleashing AI for Accelerated Scientific Research (2024.emnlp-demo)

Copied to clipboard

Challenge: Global scientific publications are growing annually by about 4%-5% (Pinedo et al., 2024).
Approach: They introduce an AI-assisted platform that answers diverse questions from researchers using Retrieval-Augmented Generation (RAG) they develop various tools to understand queries, search from the scientific literature, filter retrieved information, provide accurate and comprehensive answers, and self-refine answers.
Outcome: OpenResearcher is built on Retrieval-Augmented Generation (RAG) to integrate Large Language Models (LLMs) with up-to-date, domain-specific knowledge.
SLIM: Let LLM Learn More and Forget Less with Soft LoRA and Identity Mixture (2025.naacl-long)

Copied to clipboard

Challenge: balancing the training budget, downstream performance, and general capabilities of large language models remains a challenge in many applications.
Approach: They propose a mixture of expert framework based on Soft LoRA and Identity Mixture . SLIM allows dynamic routing between LoRA adapters and identity layers .
Outcome: The proposed framework reduces training cost while maintaining general capabilities . it can be open-sourced upon publication.
TalkLoRA: Communication-Aware Mixture of Low-Rank Adaptation for Large Language Models (2026.acl-long)

Copied to clipboard

Challenge: Existing LoRA methods assume that experts operate independently, leading to unstable routing, expert dominance.
Approach: They propose a communication-aware MoELoRA framework that relaxes this assumption by introducing expert-level communication prior to routing.
Outcome: The proposed framework outperforms vanilla LoRA and MoELoRA on diverse language understanding tasks while maintaining expert dominance.
DICP: Deep In-Context Prompt for Event Causality Identification (2025.findings-emnlp)

Copied to clipboard

Challenge: Existing prompt-learning-based methods concatenate in-context examples only at the input layer, limiting the model’s ability to capture abstract semantic cues necessary for identifying complex causal relationships.
Approach: They propose a model that injects in-context examples into the deeper layer of a pre-trained language model (PLM) this model leverages hierarchical semantic representations formed in deeper layers, thereby enhancing its capacity to learn high-level causal abstractions.
Outcome: The proposed model improves on two widely used datasets and shows that it can learn high-level causal abstractions.
Learn to Adapt for Generalized Zero-Shot Text Classification (2022.acl-long)

Copied to clipboard

Challenge: Existing methods for generalized zero-shot text classification generalize poorly since the learned parameters are only optimal for seen classes rather than for both classes.
Approach: They propose a network that trains an adaptive classifier by using both seen and virtual unseen classes to simulate a generalized zero-shot learning scenario.
Outcome: The proposed model outperforms several previous approaches on five text classification datasets.
Self-Demos: Eliciting Out-of-Demonstration Generalizability in Large Language Models (2024.findings-naacl)

Copied to clipboard

Challenge: Existing methods that rely on limited demos and out-of-demonstration (OOD) queries fail when faced with out- of-demotion queries.
Approach: They propose a query-aware prompting method that elicits the inherent generalizability of large language models by query-based demo generation.
Outcome: The proposed method outperforms state-of-the-art methods in the OOD setting and two public math benchmarks.
DenseLoRA: Dense Low-Rank Adaptation of Large Language Models (2025.acl-long)

Copied to clipboard

Challenge: Low-rank adaptation (LoRA) is an efficient approach for adapting large language models (LLMs) but many of the weights in these matrices are redundant, leading to inefficiencies in parameter utilization.
Approach: They propose a low-rank adaptation approach that fine-tunes two low-ranked matrices and adapts them through a dense low-Rank matrix, improving parameter utilization and adaptation efficiency.
Outcome: The proposed approach achieves 83.8% accuracy with only 0.01% of trainable parameters compared to LoRA's 80.8% with 0.70% of trainability parameters on LLaMA3-8B.
GraphLoRA: Structure-Aware Low-Rank Adaptation for Large Language Model Recommendation (2026.findings-acl)

Copied to clipboard

Challenge: Existing methods for translating collaborative information into textual prompts or injecting pre-trained embeddings into the LLM treat structural information as static input and fail to capture high-order relational dependencies.
Approach: They propose a framework that generalizes low-rank adaptation from independent to structure-aware propagation by embedding a trainable graph message-passing network within the low-ranked adaptation pathway.
Outcome: Experiments on multiple benchmarks show that GraphLoRA outperforms state-of-the-art recommendation methods and achieves superior generalization.
Distill Visual Chart Reasoning Ability from LLMs to MLLMs (2025.findings-emnlp)

Copied to clipboard

Challenge: a new method for generating chart annotations is proposed to improve visual reasoning in multimodal large language models.
Approach: They propose a code-as-intermediary translation method for distilling visual reasoning abilities from LLMs to MLLMs.
Outcome: The proposed method is cost-effective, efficient and scalable.
Revisiting Audio-language Pretraining for Learning General-purpose Audio Representation (2026.acl-long)

Copied to clipboard

Challenge: Existing methods for learning general-purpose audio representations are limited in scope and coverage of audio attributes.
Approach: They propose to use a 10.7M caption dataset to compare ALP with captioning . they find that contrastive learning yields competitive, transferable representations .
Outcome: The proposed model yields competitive, transferable representations, while captioning exhibits better scalability.
Show, Write, and Retrieve: Entity-aware Article Generation and Retrieval (2023.findings-emnlp)

Copied to clipboard

Challenge: Prior work typically encodes all tokens in articles using pre-trained language models, however, many named entities are difficult to accurately recognize and predict by language models.
Approach: They propose an ENtity-aware article GeneratIoN and rEtrieval framework to explicitly incorporate named entities into language models.
Outcome: The proposed framework can boost article generation and retrieval performance, with a 4-5 perplexity improvement in article generation, and a 3-4% boost in recall@1 in article retrieval.
DR-Arena: an Automated Evaluation Framework for Deep Research Agents (2026.acl-long)

Copied to clipboard

Challenge: Existing benchmarks for evaluating deep research capabilities rely on static datasets.
Approach: They propose a fully automated evaluation framework that pushes DR agents to their capability limits through dynamic investigation.
Outcome: DR-Arena achieves a Spearman correlation of 0.94 with the LMSYS Search Arena leaderboard.

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