Retrieval Augmented Models

Retrieval Augmented Models use an external source of data at query / inference time to augment the model.

Some benefits include:

  1. Reduce the need to encode knowledge in model parameters
  2. Can be updated on the fly without retraining
  3. Scale better than large models
  4. Can be audited and corrected without retraining

Vision

Papers

NLP

Papers

Ideas

  • Knowledge Graph Augmented (GNNs / Transformers to aggregate neighborhood)
  • Retrieval Augmented Instance Recognition (Few Shot)