가제: Bridging Modality and Preference Gaps in Vision-Language Models: Methods and Benchmarks for Human-Aligned Multimodal Learning


1. Introduction

목적

핵심 문제

논문 contribution 요약


2. 관련 연구 (Background)