Approaches to improve recommender systems/collaborative filtering

Part 1 (2020)
1

I’ve been learning recommender systems/collaborative filtering and I would like to know if there are any ways to improve model performance?

I understand that some approaches are

  • Weight decay (reduce overfitting)
  • Generate + introduce synthetic data (good but may introduce bias)

I’m currently digging through the fastai internals to see if something could be manipulated.

Any ideas/thoughts would be greatly appreciated :slight_smile: !