De-amplifying Bias from Differential Privacy in Language Model Fine-tuning.
Sanjari SrivastavaPiotr MardzielZhikhun ZhangArchana AhlawatAnupam DattaJohn C. MitchellPublished in: CoRR (2024)
Keyphrases
- language model
- fine tuning
- differential privacy
- differentially private
- language modeling
- privacy preserving
- n gram
- probabilistic model
- document retrieval
- information retrieval
- test collection
- personal information
- privacy preservation
- ad hoc information retrieval
- retrieval model
- query expansion
- data sharing
- mixture model
- search logs
- data privacy
- bayesian inference
- translation model
- knn
- search engine