Preprint
Unintended Memorization in Large ASR Models, and How to Mitigate It
Lun Wang et al., arXiv, 2023
Read the originalRead on October 17, 2026
- What it is
- A 2023 preprint from Google researchers that measures how much large speech recognition models memorize their training examples, and that tests ways to reduce it. It starts from the general finding that neural networks can memorize training data and so raise privacy concerns.
- How much weight it carries
- A preprint from the company that builds large speech models. The method measures memorization with specially built examples, so it shows that memorization can be measured in those models, not how much a particular commercial product memorizes.
Facts taken from this source
- The 2023 Google study states that neural networks can unintentionally memorize their training examples, which causes privacy concerns, and it sets out to measure this in large speech recognition models.
“It is well-known that neural networks can unintentionally memorize their training examples, causing privacy concerns.” (Abstract)