Awesome-Personalized-RAG-Agent

📘 Overview

This is the official repository of the paper "A Survey of Personalization: From RAG to Agent", arXiv.

🎯 Our repository Awesome-Personalized-RAG-Agent is a curated collection of papers, resources, benchmarks, and datasets focused on Personalized Retrieval-Augmented Generation (RAG) and personalized agentic RAG system.

Personalization has become a cornerstone in modern AI systems, enabling customized interactions that reflect individual user preferences, contexts, and goals. Recent research has increasingly explored RAG frameworks and their evolution into agent-based architectures, aiming to improve user alignment and satisfaction.

This repository systematically categorizes personalization across the three core stages of RAG:

• Pre-retrieval (e.g., query rewriting and expansion),
• Retrieval (e.g., indexing, personalized reranking),
• Generation (e.g., using explicit or implicit user signals).

Beyond traditional RAG pipelines, we extend the scope to Personalized LLM Agents—systems enhanced with agentic functionalities such as dynamic user modeling, personalized planning, memory integration, and autonomous behavior.

🔥 News

• [2025, Apr 15]:  🚀🚀 Our paper is now available on arXiv, and the reading list is on GitHub Repo.

🧭 Table of Contents

• 📃 Papers
  • 1. Pre-retrieval
    • 🔄 1.1 Query Rewriting
    • ➕ 1.2 Query Expansion
    • 🛠️ 1.3 Other Query-related
  • 2. Retrieval
    • 🗂️ 2.1 Indexing
    • 🔍 2.2 Retrieve
    • 🧹 2.3 Post-Retrieve
  • 3. Generation
    • 🎯 3.1 Generation from Explicit Preference
    • 🕵️ 3.2 Generation from Implicit Preference
  • 4. Agentic RAG
    • 🧠 4.1 Understanding
    • 🗺️ 4.2 Planing and Execution
    • ✍️ 4.3 Generation
• 📚 Datasets and Evaluation
• 🔗 Related Surveys and Repositories
  • 📚 Surveys
  • 📁 Repositories
• 🚀 Contributing
• 📌 Citation

📃 Papers

1. Pre-retrieval

🔄 1.1 Query Rewriting

Name	Title	Personalized presentation	Publication	Paper Link	Code Link
Least-to-most Prompting	Least-to-Most Prompting Enables Complex Reasoning in Large Language Models	Split by sub-query terms in different questions	ICLR 2023	[Link]
	Personalized Search-based Query Rewrite System for Conversational AI	Build a personalized index for each user	ACL 2021	[Link]
Agent4Ranking	Agent4Ranking: Semantic Robust Ranking via Personalized Query Rewriting Using Multi-agent LLM	Use agents for efficient query rewriting.	arXiv 2023	[Link]
Least-to-most Prompting	Query Rewriting in TaoBao Search	A learning enhanced architecture based on “query retrieval−semantic relevance”.	CIKM 2022	[Link]
CLE-QR	Learning to rewrite prompts for personalized text generation	Multistage framework for personalized rewrites.	WWW 2024	[Link]
CGF	CGF: Constrained Generation Framework for Query Rewriting in Conversational AI	Personalized prompt rewriting by using an LLM agent.	ACL 2022	[Link]
	RL-based Query Rewriting with Distilled LLM for online E-Commerce Systems	Student model to rewrite query.	arXiv 2025	[Link]
CoPS	Cognitive Personalized Search Integrating Large Language Models with an Efficient Memory Mechanism	Personalized query intent.	WWW 2024	[Link]
BASES	BASES: Large-scale Web Search User Simulation with Large Language Model based Agents	User simulation agent.	arXiv 2024	[Link]
ERAGent	ERAGent: Enhancing Retrieval-Augmented Language Models with Improved Accuracy, Efficiency, and Personalization	Collaorative module for query rewrite.	arXiv 2024	[Link]
PEARL	PEARL: Personalizing Large Language Model Writing Assistants with Generation-Calibrated Retrievers	Personlaized LLM for query write.	arXiv 2024	[Link]
FIG	Graph Meets LLM: A Novel Approach to Collaborative Filtering for Robust Conversational Understanding	Graph-based methods with LLMs to query rewrite	arXiv 2023	[Link]

➕ 1.2 Query Expansion

Name	Title	Personalized presentation	Publication	Paper Link	Code Link
PNQES	Personalized Optimal Search in Local Query Expansion	Search history, Latent semantic indexing	ACLCLP 2006	[Link]
	Exploiting social relations for query expansion and result ranking	Friendship, Tagging	ICDE Workshop 2008	[Link]
Gossple	Toward personalized query expansion	Calculating the distance between users, constructing a personalized network that connects users with similar interests. Extracting tags	EuroSys Workshop 2009	[Link]
	Social tagging in query expansion: A new way for personalized web search	User Interest Tagging	CSE 2009	[Link]
SoQuES	personalized social query expansion using social bookmarking systems	Extract user tag behavior to build personalized user profiles.	SIGIR 2011	[Link]
	Improving search via personalized query expansion using social media	Tagging	Information retrieval 2012	[Link]
	Axiomatic term-based personalized query expansion using bookmarking system	Bookmarking, Social Network	International Conference on Database and Expert Systems Applications 2016	[Link]
WE-LM	personalized query expansion utilizing multi-relationalsocial data	Tagging	SMAP Workshop 2017	[Link]
PSQE	personalized social query expansion using social annotations	User Interest Tagging	Transactions on Large-Scale Data-and Knowledge-Centered Systems XL 2019	[Link]
PQEWC	Personalized Query Expansion with Contextual Word Embeddings	Employing topic modeling on user texts and dynamically selecting relevant words	TOIS 2023	[Link]

🛠️ 1.3 Other Query-related

Name	Title	Personalized presentation	Publication	Paper Link	Code Link
PSQE	PSQE: Personalized Semantic Query Expansion for user-centric query disambiguation	Leveraging synthetic user profiles built from Wikipedia articles, training word2vec embeddings on these profiles		[Link]
Bobo	Utilizing user-input contextual terms for query disambiguation	contextual terms	Coling 2010	[Link]
	Personalized Query Auto-Completion Through a Lightweight Representation of the User Context	Learning embeddings from the user’s recent queries	arXiv 2019	[Link]

2. Retrieval

🗂️ 2.1 Indexing

Name	Title	Personalized presentation	Publication	Paper Link	Code Link
Pearl	Pearl: Personalizing large language model writing assistants with generation-calibrated retrievers	Personalized Indexing	ACL 2024	[Link]
KG-Retrieve	KG-Retriever: Efficient Knowledge Indexing for Retrieval-Augmented Large Language Models	Hierarchical Index Graph	arXiv 2024	[Link]
EMG-RAG	Crafting Personalized Agents through Retrieval-Augmented Generation on Editable Memory Graphs	Editable Knowledge Graph for Indexing	EMNLP 2024	[Link]
PGraphRAG	Personalized Graph-Based Retrieval for Large Language Models	User-centric Knowledge Graphs for Personalized Indexing	arXiv 2025	[Link]	[Link]

🔍 2.2 Retrieve

Name	Title	Personalized presentation	Publication	Paper Link	Code Link
	Optimization Methods for Personalizing Large Language Models through Retrieval Augmentation	Gradients based on personalized scores	SIGIR 2024	[Link]
MeMemo	MeMemo: On-device Retrieval Augmentation for Private and Personalized Text Generation	Privacy Protection	SIGIR 2024 (short)	[Link]	[Link]
LAPS	Doing Personal LAPS: LLM-Augmented Dialogue Construction for Personalized Multi-Session Conversational Search	Personalized Dialogue	SIGIR 2024	[Link]	[Link]
	Partner Matters! An Empirical Study on Fusing Personas for Personalized Response Selection in Retrieval-Based Chatbots	Personalized Dialogue	SIGIR 2021	[Link]	[Link]
ERRA	Explainable Recommendation with Personalized Review Retrieval and Aspect Learning	Personalized Recommendation	ACL 2023	[Link]	[Link]
	RECAP: Retrieval-Enhanced Context-Aware Prefix Encoder for Personalized Dialogue Response Generation	Personalized Dialogue	ACL 2023	[Link]	[Link]
HEART	HEART-felt Narratives: Tracing Empathy and Narrative Style in Personal Stories with LLMs	Personalized Writing Style	EMNLP 2024	[Link]	[Link]
OPPU	Democratizing Large Language Models via Personalized Parameter-Efficient Fine-tuning	Personalized Parameter Fine-tuning	EMNLP 2024	[Link]	[Link]
LAPDOG	Learning Retrieval Augmentation for Personalized Dialogue Generation	Personalized Dialogue	EMNLP 2023	[Link]	[Link]
UniMP	Towards Unified Multi-Modal Personalization: Large Vision-Language Models for Generative Recommendation and Beyond	Personalized Recommendation	ICLR 2024	[Link]
	Personalized Language Generation via Bayesian Metric Augmented Retrieval	Personalized Retrieval	arXiv	[Link]
	Leveraging Similar Users for Personalized Language Modeling with Limited Data	Personalized Retrieval	ACL 2022	[Link]
UIA	A Personalized Dense Retrieval Framework for Unified Information Access	Personalized Retrieval	SIGIR 2023	[Link]	[Link]
XPERT	Personalized Retrieval over Millions of Items	Personalized Retrieval	SIGIR 2023	[Link]	[Link]
DPSR	Towards personalized and semantic retrieval: An end-to-end solution for e-commerce search via embedding learning	Personalized Retrieval	SIGIR 2020	[Link]
PersonalTM	PersonalTM: Transformer Memory for Personalized Retrieval	Personalized Retrieval	SIGIR 2023 (short)	[Link]
	A zero attention model for personalized product search	Personalized Search	CIKM 2019	[Link]
RTM	Learning a Fine-Grained Review-based Transformer Model for Personalized Product Search	Personalized Search	SIGIR 2021	[Link]	[Link]

🧹 2.3 Post-Retrieve

Name	Title	Personalized presentation	Publication	Paper Link	Code Link
PersonaRAG	PersonaRAG: Enhancing Retrieval-Augmented Generation Systems with User-Centric Agents	User-centric Agent for Ranking	arXiv 2024	Link	Link
	Improving RAG with LoRA Finetuning for Persona Text Generation	Personalized Re-ranking		Link
UniMS-RAG	UniMS-RAG: A Unified Multi-source Retrieval-Augmented Generation for Personalized Dialogue Systems	Personalized Re-ranking	arXiv 2024	Link
	Learning to Rank for Multiple Retrieval-Augmented Models through Iterative Utility Maximization	Personalized Ranking	arXiv 2024	Link	Link
	Rehearse With User: Personalized Opinion Summarization via Role-Playing Based on Large Language Models	Personalized Opinion Summarization	arXiv 2025	Link
AutoCompressors	Adapting Language Models to Compress Contexts	Contextual Embedding Compression	EMNLP 2023	Link	Link
FIT-RAG	FIT-RAG: Black-Box RAG with Factual Information and Token Reduction	Tokens Minization within RAG	ToIS 2025	Link

3. Generation

🎯 3.1 Generation from Explicit Preference

Name	Title	Personalized presentation	Publication	Paper Link	Code Link
P2	Evaluating and inducing personality in pre-trained language models	Role Playing	NeurIPS 2023	[Link]	[Link]
OpinionQA	Whose opinions do language models reflect?	Role Playing	ICML 2023	[Link]	[Link]
Character Profiling	Evaluating Character Understanding of Large Language Models via Character Profiling from Fictional Works	Role Playing	ICML 2023	[Link]	[Link]
	Do LLMs Understand User Preferences? Evaluating LLMs On User Rating Prediction	Personalized Recommendation	arXiv	[Link]
Cue-CoT	Cue-CoT: Chain-of-thought prompting for responding to in-depth dialogue questions with LLMs	Personalized Dialogue	EMNLP 2023	[Link]	[Link]
TICL	Tuning-Free Personalized Alignment via Trial-Error-Explain In-Context Learning	Personalized Text Generation	arXiv	[Link]	[Link]
GPG	Guided Profile Generation Improves Personalization with LLMs	Personalized Text Generation	arXiv	[Link]
	Integrating Summarization and Retrieval for Enhanced Personalization via Large Language Models	Personalized Text Generation	arXiv	[Link]
LLMTreeRec	LLMTreeRec: Unleashing the Power of Large Language Models for Cold-Start Recommendations	Personalized Recommendation	COLING 2025	[Link]	[Link]
Matryoshka	MATRYOSHKA: Learning To Drive Black-Box LLMS With LLMS	Personalized Text Generation	arXiv	[Link]
	Learning to rewrite prompts for personalized text generation	Personalized Text Generation	WWW 2024	[Link]
RecGPT	RecGPT: Generative Pre-training for Text-based Recommendation	Personalized Recommendation	ACL 2024	[Link]	[Link]
PEPLER-D	Personalized prompt learning for explainable recommendation	Personalized Recommendation	TOIS 2023	[Link]	[Link]
SGPT	Unlocking the potential of prompt-tuning in bridging generalized and personalized federated learning	Personalized Federated Learning	CVPR 2024	[Link]	[Link]
PFCL	Personalized federated continual learning via multi-granularity prompt	Personalized Federated Learning	KDD 2024	[Link]	[Link]

🕵️ 3.2 Generation from Implicit Preference

Name	Title	Personalized presentation	Publication	Paper Link	Code Link
PLoRA	Personalized LoRA for Human-Centered Text Understanding	Personalized Text Understanding	AAAI 2024	[Link]	[Link]
LM-P	Personalized Large Language Models	Personalized Fine-tuning	SENTIRE 2024 (ICDM Workshop)	[Link]	[Link]
MiLP	Personalized LLM Response Generation with Parameterized User Memory Injection	Personalized Text Generation	arXiv	[Link]	[Link]
OPPU	Democratizing Large Language Models via Personalized Parameter-Efficient Fine-tuning	Personalized Text Generation	EMNLP 2024	[Link]	[Link]
PER-PCS	PERSONALIZED PIECES: Efficient Personalized Large Language Models through Collaborative Efforts	Personalized Text Generation	EMNLP 2024	[Link]	[Link]
Review-LLM	Review-LLM: Harnessing Large Language Models for Personalized Review Generation	Personalized Review Generation	arXiv	[Link]
UserIdentifier	UserIdentifier: Implicit User Representations for Simple and Effective Personalized Sentiment Analysis	Personalized Text Understanding	NAACL 2022	[Link]
UserAdapter	UserAdapter: Few-Shot User Learning in Sentiment Analysis	Personalized Text Understanding	ACL Fingdings 2021	[Link]
HYDRA	HYDRA: Model Factorization Framework for Black-Box LLM Personalization	Personalized Reranking and Generation	NeurIPS 2024	[Link]	[Link]
PocketLLM	PocketLLM: Enabling On-Device Fine-Tuning for Personalized LLMs	Personalized Text Generation	PrivateNLP 2024 (ACL Workshop)	[Link]
CoGenesis	CoGenesis: A Framework Collaborating Large and Small Language Models for Secure Context-Aware Instruction Following	Personalized Text Generation	ACl 2024	[Link]	[Link]
P-RLHF	P-RLHF: Personalized Language Modeling from Personalized Human Feedback	Personalized Text Generation	arXiv	[Link]	[Link]
P-SOUPS	Personalized Soups: Personalized Large Language Model Alignment via Post-hoc Parameter Merging	Personalized Text Generation	Adaptive Foundation Models 2024 (NeurIPS 2024)	[Link]	[Link]
PAD	PAD: Personalized Alignment of LLMs at Decoding-Time	Personalized Text Generation	ICLR 2025	[Link]	[Link]
REST-PG	Reasoning-enhanced self-training for long-form personalized Text Generation	Personalized Text Generation	arXiv	[Link]
	Optimization Methods for Personalizing Large Language Models through Retrieval Augmentation	Personalized Retrieval and Generation	SIGIR 2024	[Link]
RewriterSlRl	Learning to Rewrite Prompts for Personalized Text Generation	Personalized Text Generation	WWW 2024	[Link]
	Reinforcement learning for optimizing rag for domain chatbots	Personalized Chatbot	RL+LLMs 2024 (AAAI Workshop)	[Link]

4. Agentic RAG

🧠 4.1 Understanding

Name	Title	Personalized Presentation	Publication	Paper Link	Code Link
	Penetrative AI: Making LLMs Comprehend the Physical World	User interaction with physical-world data via sensors	ACL Findings 2024	Link	Link
	Conversational Health Agents: A Personalized LLM-Powered Agent Framework	Personalized healthcare support via health data and knowledge base	arXiv 2024	Link	Link
RoleLLM	RoleLLM: Benchmarking, eliciting, and enhancing role-playing abilities of large language models	LLMs role-playing characters	arXiv 2024	Link
Character-llm	Character-llm: A trainable agent for role-playing	LLM character simulation with memory/personality	EMNLP 2023	Link	Link
InCharacter	Evaluating personality fidelity in role-playing agents through psychological interviews	Psychological evaluation of role-play agents	ACL 2024	Link	Link
Socialbench	Socialbench: Sociality evaluation of role-playing conversational agents	Benchmarking social interaction ability in role-playing	ACL Findings 2024	Link	Link
MMrole	Mmrole: A comprehensive framework for developing and evaluating multimodal role-playing agents	Developing and Evaluating Multimodal Role-Playing Agents	ICLR 2025	Link	Link
	Enhancing role-playing language models with personality-indicative data	Leveraging psychological scale questions for agent personality modelling	EMNLP Findings 2024	Link	Link
	Enabling Conversational Interaction with Mobile UI using Large Language Models	Mobile UI interaction with LLMs	CHI 2023	Link
CharacterEval	A chinese benchmark for role-playing conversational agent evaluation	Chinese dataset for role-playing agent evaluation	arXiv 2024	Link	Link
Neeko	Neeko: Leveraging dynamic lora for efficient multi-character role-playing agent	Efficient multi-character simulation	EMNLP 2024	Link	Link

🗺️ 4.2 Planing and Execution

Name	Title	Personalized Presentation	Publication	Paper Link	Code Link
EMG-RAG	Crafting Personalized Agents through Retrieval-Augmented Generation on Editable Memory Graphs	Personalized assistant using user memory	EMNLP 2024	Link
	Generative agents: Interactive simulacra of human behavior	Simulation of social and individual behaviors in Stanford Town	UIST 2023	Link
	Conversational Health Agents: A Personalized LLM-Powered Agent Framework	Personalized healthcare support via health data and knowledge base	arXiv 2024	Link	Link
	User Behavior Simulation with Large Language Model based Agents	Simulation of user behavior in recommender systems	arXiv 2024	Link
	Personal Large Language Model Agents: A Case Study on Tailored Travel Planning	Tailored travel based on user preferences	EMNLP 2024	Link
PersonalWAB	Large Language Models Empowered Personalized Web Agents	User profile + web action optimization	WWW 2025 Oral	Link
	Voyager: An Open-Ended Embodied Agent with Large Language Models	Minecraft agent simulation and skill learning	TMLR 2024	Link	Link
	MEMORYLLM: Towards Self-Updatable Large Language Models	Self-evolution via memory update	ICML 2024	Link
	Bootstrap Your Own Skills: Learning to Solve New Tasks with Large Language Model Guidance	Robotic arm learning to manipulate objects	CoRL 2023	Link	Link
	Enabling Conversational Interaction with Mobile UI using Large Language Models	Mobile UI interaction with LLMs	CHI 2023	Link
	Penetrative AI: Making LLMs Comprehend the Physical World	User interaction with physical-world data via sensors	ACL Findings 2024	Link	Link
	Language models as zero-shot planners: Extracting actionable knowledge for embodied agents	Task planning in virtual environments	ICML 2022	Link	Link
MetaGPT	Metagpt: Meta programming for multi-agent collaborative framework	Multi-agent collaboration for complex tasks	ICLR 2024	Link
OKR-Agent	Agents meet okr: An object and key results driven agent system with hierarchical self-collaboration and self-evaluation	Role-assigned agents solving creative tasks	arXiv 2023	Link	Link

✍️ 4.3 Generation

Name	Title	Personalized Presentation	Publication	Paper Link	Code Link
Character-llm	Character-llm: A trainable agent for role-playing	LLM character simulation with memory/personality	EMNLP 2023	Link	Link
	Investigating the Personality Consistency in Quantized Role-Playing Dialogue Agents	Robustness and consistency in constrained setups	EMNLP 2024	Link
MMrole	Mmrole: A comprehensive framework for developing and evaluating multimodal role-playing agents	Developing and Evaluating Multimodal Role-Playing Agents	ICLR 2025	Link	Link
RoleLLM	RoleLLM: Benchmarking, eliciting, and enhancing role-playing abilities of large language models	LLMs role-playing characters	arXiv 2024	Link
	Enhancing role-playing language models with personality-indicative data	Personality-focused character simulation	EMNLP 2024 Findings	Link	Link
InCharacter	Evaluating personality fidelity in role-playing agents through psychological interviews	Psychological evaluation of role-play agents	ACL 2024	Link	Link
Socialbench	Socialbench: Sociality evaluation of role-playing conversational agents	Benchmarking social interaction ability in role-playing	ACL Findings 2024	Link	Link

📚 Datasets and Evaluation

Field	Dataset	Metrics	Link
Query Rewriting	SCAN	Accuracy	Link
Query Rewriting	Robust04	Accuracy	Link
Query Rewriting	Avocado Research Email Collection	BLEU, ROUGE	Link
Query Rewriting	Amazon Review	BLEU, ROUGE	Link
Query Rewriting	Reddit Comments	BLEU, ROUGE	Link
Query Rewriting	Amazon ESCI Dataset	EM, ROUGE-L, XEntropy	Link
Query Rewriting	AOL	MAP, MRR, P@1	Link
Query Rewriting	WARRIORS	MRR, NDCG	Link
Query Rewriting	AITA WORKSM	Macro-F1, BS-F1	Link
Query Rewriting	PIP	PMS, Image-Align, ROUGE	Link
Query Expansion	Personalized Results Re-Ranking	MAP, MRR, NDCG, RBP	Link
Query Expansion	del.icio.us	Precision, MAP, MRR, Recall	Link
Query Expansion	Flickr	Precision	Link
Query Expansion	CiteULike	Recall, MAP, MRR	Link
Query Expansion	LRDP	Precision, Recall, F1	Link
Query Expansion	Delicious	MAP, MRR	Link
Query Expansion	Flickr	MAP, MRR	Link
Query Expansion	Bibsonomy	MAP, Precision, PQEC, Prof-overlap	Link
Other Query-related	Wikipedia	Precision, Recall	Link
Retrieval / Generation	TOPDIAL	BLEU, F1, Success Rate	Link
Retrieval / Generation	LiveChat	Recall, MRR	Link
Retrieval / Generation	PersonalityEvd	Accuracy, Fluency, Coherence, Plausibility	Link
Retrieval / Generation	Pchatbot	BLEU, ROUGE, Distinct, MRR	Link
Retrieval / Generation	DuLemon	Perplexity, BLEU, Accuracy, Precision, Recall, F1	Link
Retrieval / Generation	PersonalityEdit	ES, DD, Accuracy, TPEI, PAE	Link
Generation	LaMP	Accuracy, F1, MAE, RMSE, ROUGE	Link
Generation	LongLaMP	Accuracy, F1, MAE, RMSE, ROUGE	Link
Generation	PGraphRAG	ROUGE, METEOR, MAE, RMSE	Link
Generation	AmazonQA Products	ROUGE, Persona-F1	Link
Generation	Reddit	ROUGE, Persona-F1	Link
Generation	MedicalDialogue	ROUGE, Persona-F1	Link
Generation	Personalized-gen	Mean Success Rate, Median Relative Improvements, Fluency	Link

🔗 Related Surveys and Repositories

📚 Surveys

🧠 Personalized LLMs

• When Large Language Models Meet Personalization: Perspectives of Challenges and Opportunities
• Personalization of Large Language Models: A Survey
• A Survey of Personalized Large Language Models: Progress and Future Directions

🎭 Personalized Role-Playing

• From Persona to Personalization: A Survey on Role-Playing Language Agents
• Two Tales of Persona in LLMs: A Survey of Role-Playing and Personalization

📁 Repositories

• Awesome Personalized Large Language Models (PLLMs)
  A curated list of resources on personalized large language models.

• PersonaLLM Survey
  Companion repository for the PersonaLLM survey, covering role-playing and personalization.

• Awesome Personalized LLM
  A collection of papers and tools focused on personalized LLM development.

🚀 Contributing

We sincerely welcome you to contribute to this repository!

🙌 Whether you're adding new papers or datasets/benchmarks, fixing bugs, improving the documentation, or suggesting ideas, every bit of help is appreciated.

📌 Citation

If you find this repository useful in your research, please consider citing our paper:

@article{li2025survey,
  title={A Survey of Personalization: From RAG to Agent},
  author={Li, Xiaopeng and Jia, Pengyue and Xu, Derong and Wen, Yi and Zhang, Yingyi and Zhang, Wenlin and Wang, Wanyu and Wang, Yichao and Du, Zhaocheng and Li, Xiangyang and others},
  journal={arXiv preprint arXiv:2504.10147},
  year={2025}
}