Personalized Pieces: Efficient Personalized Large Language Models through Collaborative Efforts

TL;DR

This paper introduces Personalized Pieces (Per-Pcs), a collaborative framework for efficient, privacy-preserving personalization of large language models through sharing and assembling PEFT components, achieving high performance with lower resource use.

Contribution

The paper presents a novel collaborative PEFT sharing framework, enabling efficient, privacy-preserving personalization of LLMs with reduced storage and computation.

Findings

Per-Pcs outperforms non-personalized and PEFT retrieval baselines.

Achieves comparable performance to OPPU with lower resource consumption.

Robust across different sharer counts and sharing strategies.

Abstract

Personalized large language models (LLMs) aim to tailor interactions, content, and recommendations to individual user preferences. While parameter-efficient fine-tuning (PEFT) methods excel in performance and generalization, they are costly and limit communal benefits when used individually. To this end, we introduce Personalized Pieces (Per-Pcs), a framework that allows users to safely share and assemble personalized PEFT efficiently with collaborative efforts. Per-Pcs involves selecting sharers, breaking their PEFT into pieces, and training gates for each piece. These pieces are added to a pool, from which target users can select and assemble personalized PEFT using their history data. This approach preserves privacy and enables fine-grained user modeling without excessive storage and computation demands. Experimental results show Per-Pcs outperforms non-personalized and PEFT retrieval baselines, offering performance comparable to OPPU with significantly lower resource use across six tasks. Further analysis highlights Per-Pcs's robustness concerning sharer count and selection strategy, pieces sharing ratio, and scalability in computation time and storage space. Per-Pcs's modularity promotes safe sharing, making LLM personalization more efficient, effective, and widely accessible through collaborative efforts.