Answer: Transfer learning

## Explanation **Transfer learning** is the most efficient approach in this scenario because: - The hospital already has a trained model for pneumonia detection from chest X-rays - Both tasks (pneumonia detection and lung cancer detection) use the same: - Base architecture - Dataset type (chest X-rays) - Domain (medical imaging) **Why transfer learning works best:** 1. **Reuses learned features**: The model has already learned relevant features from chest X-rays that are useful for both pneumonia and lung cancer detection 2. **Faster training**: Only the final layers need to be retrained or fine-tuned, significantly reducing training time 3. **Less data required**: Transfer learning typically requires less labeled data for the new task 4. **Better performance**: Leveraging pre-trained weights often leads to better performance compared to training from scratch **Why other options are less efficient:** - **Unsupervised learning**: Would require learning patterns from scratch without labels - **Self-supervised learning**: Involves creating pseudo-labels, which is unnecessary when you already have a relevant pre-trained model - **Reinforcement learning**: Not suitable for this type of classification task; designed for sequential decision-making problems Transfer learning allows the hospital to build upon their existing investment in the pneumonia detection model, making the adaptation to lung cancer detection much more efficient and effective.

Author: Ritesh Yadav