GORACS: Group-level Optimal Transport-guided Coreset Selection for LLM-based Recommender Systems

2025/06/04 by Tiehua Mei, Mei, Tiehua, Hengrui Chen +7 · 2 citations
Computer Science · #Topic Modeling #Recommender Systems and Techniques #Multimodal Machine Learning Applications

paper · pdf · doi:10.48550/arxiv.2506.04015

Abstract

Although large language models (LLMs) have shown great potential in recommender systems, the prohibitive computational costs for fine-tuning LLMs on entire datasets hinder their successful deployment in real-world scenarios. To develop affordable and effective LLM-based recommender systems, we focus on the task of coreset selection which identifies a small subset of fine-tuning data to optimize the test loss, thereby facilitating efficient LLMs' fine-tuning. Although there exist some intuitive solutions of subset selection, including distribution-based and importance-based approaches, they often lead to suboptimal performance due to the misalignment with downstream fine-tuning objectives or weak generalization ability caused by individual-level sample selection. To overcome these challenges, we propose GORACS, which is a novel Group-level Optimal tRAnsport-guided Coreset Selection framework for LLM-based recommender systems. GORACS is designed based on two key principles for coreset selection: 1) selecting the subsets that minimize the test loss to align with fine-tuning objectives, and 2) enhancing model generalization through group-level data selection. Corresponding to these two principles, GORACS has two key components: 1) a Proxy Optimization Objective (POO) leveraging optimal transport and gradient information to bound the intractable test loss, thus reducing computational costs by avoiding repeated LLM retraining, and 2) a two-stage Initialization-Then-Refinement Algorithm (ITRA) for efficient group-level selection. Our extensive experiments across diverse recommendation datasets and tasks validate that GORACS significantly reduces fine-tuning costs of LLMs while achieving superior performance over the state-of-the-art baselines and full data training. The source code of GORACS are available at https://github.com/Mithas-114/GORACS.

Citations

• Speculative Coreset Selection for Task-Specific Fine-tuning
• Most Influential Subset Selection: Challenges, Promises, and Beyond
• HLLM: Enhancing Sequential Recommendations via Hierarchical Large Language Models for Item and User Modeling
• Generative Explore-Exploit: Training-free Optimization of Generative Recommender Systems using LLM Optimizers
• Text-like Encoding of Collaborative Information in Large Language Models for Recommendation
• Get more for less: Principled Data Selection for Warming Up Fine-Tuning in LLMs
• Aligning Large Language Models with Recommendation Knowledge
• Large Language Models Enhanced Collaborative Filtering
• Item-side Fairness of Large Language Model-based Recommendation System
• LESS: Selecting Influential Data for Targeted Instruction Tuning
• A Survey on Data Selection for LLM Instruction Tuning
• A Survey on Data-Centric Recommender Systems
• Data-efficient Fine-tuning for LLM-based Recommendation
• LLMRec: Large Language Models with Graph Augmentation for Recommendation
• Dataset Condensation for Recommendation
• A Bi-Step Grounding Paradigm for Large Language Models in Recommendation Systems
• LLM-Rec: Personalized Recommendation via Prompting Large Language Models
• Fairness and Diversity in Recommender Systems: A Survey
• A Survey on Large Language Models for Recommendation
• Farewell to Aimless Large-scale Pretraining: Influential Subset Selection for Language Model
• INGENIOUS: Using Informative Data Subsets for Efficient Pre-Training of Language Models
• LAVA: Data Valuation without Pre-Specified Learning Algorithms
• Is ChatGPT a Good Recommender? A Preliminary Study
• Chat-REC: Towards Interactive and Explainable LLMs-Augmented Recommender System
• TRAK: Attributing Model Behavior at Scale
• Data Selection for Language Models via Importance Resampling
• Handbook of Convergence Theorems for (Stochastic) Gradient Methods
• Coverage-centric Coreset Selection for High Pruning Rates
• Beyond neural scaling laws: beating power law scaling via data pruning
• DeepCore: A Comprehensive Library for Coreset Selection in Deep Learning
• RETRIEVE: Coreset Selection for Efficient and Robust Semi-Supervised\n Learning
• A Critical Study on Data Leakage in Recommender System Offline Evaluation
• Coresets via Bilevel Optimization for Continual Learning and Streaming
• An Empirical Study of Example Forgetting during Deep Neural Network Learning
• Self-Attentive Sequential Recommendation
• BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding
• Nonlinear Programming

Cited by

• A Survey on Generative Recommendation: Data, Model, and Tasks
• Move What Matters: Parameter-Efficient Domain Adaptation via Optimal Transport Flow for Collaborative Perception

Related

• Data-efficient Fine-tuning for LLM-based Recommendation
• Does LLM Focus on the Right Words? Mitigating Context Bias in LLM-based Recommenders
• Architecture is All You Need: Improving LLM Recommenders by Dropping the Text
• Bridging the Gap: Self-Optimized Fine-Tuning for LLM-based Recommender Systems
• OpenP5: An Open-Source Platform for Developing, Training, and Evaluating LLM-based Recommender Systems