Summary of Separable Mixture Of Low-rank Adaptation For Continual Visual Instruction Tuning, by Ziqi Wang et al.
Separable Mixture of Low-Rank Adaptation for Continual Visual Instruction Tuning
by Ziqi Wang, Chang Che, Qi Wang, Yangyang Li, Zenglin Shi, Meng Wang
First submitted to arxiv on: 21 Nov 2024
Categories
- Main: Computer Vision and Pattern Recognition (cs.CV)
- Secondary: Artificial Intelligence (cs.AI)
GrooveSquid.com Paper Summaries
GrooveSquid.com’s goal is to make artificial intelligence research accessible by summarizing AI papers in simpler terms. Each summary below covers the same AI paper, written at different levels of difficulty. The medium difficulty and low difficulty versions are original summaries written by GrooveSquid.com, while the high difficulty version is the paper’s original abstract. Feel free to learn from the version that suits you best!
| Summary difficulty | Written by | Summary |
|---|---|---|
| High | Paper authors | High Difficulty Summary Read the original abstract here |
| Medium | GrooveSquid.com (original content) | Medium Difficulty Summary The paper proposes an innovative approach to continual visual instruction tuning (CVIT) for multimodal large language models (MLLMs). It introduces the Separable Mixture of Low-Rank Adaptation (SMoLoRA) framework, which addresses dual catastrophic forgetting by separating adaptation into two modules: one for visual understanding and another for instruction following. This design enables specialized adaptation in both domains, preventing forgetting while improving performance. The paper also presents a novel CVIT benchmark that evaluates a model’s ability to generalize to unseen tasks and handle diverse instructions across various tasks. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary The research aims to improve the capabilities of MLLMs by allowing them to learn new tasks incrementally. It identifies two types of catastrophic forgetting in CVIT: not only forgetting previously learned visual understanding but also experiencing a decline in instruction following abilities as they acquire new tasks. The proposed SMoLoRA framework addresses these challenges and outperforms existing methods. |
Keywords
» Artificial intelligence » Instruction tuning » Low rank adaptation
