Model Merging in LLMs, MLLMs, and Beyond: Methods, Theories, Applications and Opportunities

TL;DR

This survey comprehensively reviews model merging techniques across various machine learning models, discussing methods, theories, applications, challenges, and future research directions without requiring raw data or heavy computation.

Contribution

It introduces a new taxonomic framework for classifying model merging methods and covers their applications in diverse domains and learning paradigms.

Findings

Provides a systematic taxonomy of model merging methods

Covers applications in LLMs, multimodal models, and subfields like continual learning

Highlights challenges and future research opportunities

Abstract

Model merging is an efficient empowerment technique in the machine learning community that does not require the collection of raw training data and does not require expensive computation. As model merging becomes increasingly prevalent across various fields, it is crucial to understand the available model merging techniques comprehensively. However, there is a significant gap in the literature regarding a systematic and thorough review of these techniques. This survey provides a comprehensive overview of model merging methods and theories, their applications in various domains and settings, and future research directions. Specifically, we first propose a new taxonomic approach that exhaustively discusses existing model merging methods. Secondly, we discuss the application of model merging techniques in large language models, multimodal large language models, and more than ten machine learning subfields, including continual learning, multi-task learning, few-shot learning, etc. Finally, we highlight the remaining challenges of model merging and discuss future research directions. A comprehensive list of papers about model merging is available at https://github.com/EnnengYang/Awesome-Model-Merging-Methods-Theories-Applications.