Summary of Attention Is Not What You Need: Revisiting Multi-instance Learning For Whole Slide Image Classification, by Xin Liu et al.
Attention Is Not What You Need: Revisiting Multi-Instance Learning for Whole Slide Image Classification
by Xin Liu, Weijia Zhang, Min-Ling Zhang
First submitted to arxiv on: 18 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 This research paper proposes a new instance-based multi-instance learning (MIL) method, FocusMIL, to improve whole slide image (WSI) classification tasks. The proposed approach addresses limitations in attention-based MIL algorithms, which often focus on irrelevant patterns and struggle with hard-to-classify instances. By synergizing standard MIL assumptions with variational inference, FocusMIL encourages models to focus on tumour morphology rather than spurious correlations. Experimental evaluations demonstrate that FocusMIL outperforms baselines on patch-level classification tasks using the Camelyon16 and TCGA-NSCLC benchmarks. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary The researchers developed a new approach for classifying whole slide images, called FocusMIL. This method helps computers focus on important details in the images instead of irrelevant things like how the image was prepared or what the surrounding tissue looks like. The old way of doing this, using attention-based multi-instance learning, wasn’t working very well because it would get distracted by these unimportant details. The new approach is simpler and more effective, and it does a better job of identifying the hard-to-classify images. |
Keywords
» Artificial intelligence » Attention » Classification » Inference
