Summary of Mora: Lora Guided Multi-modal Disease Diagnosis with Missing Modality, by Zhiyi Shi et al.
MoRA: LoRA Guided Multi-Modal Disease Diagnosis with Missing Modality
by Zhiyi Shi, Junsik Kim, Wanhua Li, Yicong Li, Hanspeter Pfister
First submitted to arxiv on: 17 Aug 2024
Categories
- Main: Computer Vision and Pattern Recognition (cs.CV)
- Secondary: Machine Learning (cs.LG)
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 In this paper, researchers develop a new method called Modality-aware Low-Rank Adaptation (MoRA) for fine-tuning multi-modal pre-trained models. These models are efficient at extracting features from different modalities but have limited applications in disease diagnosis due to issues with missing modalities and high computational requirements. MoRA addresses these challenges by projecting input data into a low intrinsic dimensionality and using modality-specific up-projections when certain modalities are missing. This approach integrates well with existing pre-trained models, improving performance without requiring significant additional computational resources. Experimental results demonstrate the superiority of MoRA over existing techniques in disease diagnosis tasks. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary MoRA is a new method for fine-tuning multi-modal pre-trained models that makes it easier to use them for disease diagnosis. These models are good at finding patterns in different types of data, but they can’t be used as much as they could because sometimes certain kinds of data are missing. This makes it hard to get accurate results. MoRA helps by changing how the model looks at the data so that it can still work even when some of the data is missing. It also doesn’t need a lot of extra computer power, which makes it faster and more efficient. |
Keywords
» Artificial intelligence » Fine tuning » Low rank adaptation » Multi modal
