In-Context Multiple Instance Learning
- Published
- Source
- arXiv
- Paper number
- 343
- Field
- Machine Learning
- arXiv ID
- 2606.06458
Key points
- A pretrained in-context learner trained on synthetic data solves MIL tasks from only a few labels.
- It classifies in a single forward pass at inference time and does not require gradient updates.
- It proposes and analyzes several synthetic data generators for bag-structured data.
- Different generators capture complementary inductive biases.
- A mixture-generator pretrained model outperforms supervised baselines on 12 MIL benchmarks.
Paper links
External research summaries. These are not HDATF publications or measured product results.