Skip to Main content Skip to Navigation
Conference papers

Assessing Unintended Memorization in Neural Discriminative Sequence Models

Abstract : Despite their success in a multitude of tasks, neural models trained on natural language have been shown to memorize the intricacies of their training data, posing a potential privacy threat. In this work, we propose a metric to quantify unintended memorization in neural dis-criminative sequence models. The proposed metric, named d-exposure (discriminative exposure), utilizes language ambiguity and classification confidence to elicit the model's propensity to memorization. Through experimental work on a named entity recognition task, we show the validity of d-exposure to measure memorization. In addition, we show that d-exposure is not a measure of overfitting as it does not increase when the model overfits.
Document type :
Conference papers
Complete list of metadatas

Cited literature [10 references]  Display  Hide  Download

https://hal.inria.fr/hal-02880581
Contributor : Zaineb Chelly Dagdia <>
Submitted on : Thursday, June 25, 2020 - 5:20:08 PM
Last modification on : Saturday, June 27, 2020 - 3:09:27 AM

File

HelaliM+20.pdf
Files produced by the author(s)

Identifiers

  • HAL Id : hal-02880581, version 1

Collections

Citation

Mossad Helali, Thomas Kleinbauer, Dietrich Klakow. Assessing Unintended Memorization in Neural Discriminative Sequence Models. 23rd International Conference on Text, Speech and Dialogue, Sep 2020, Brno, Czech Republic. ⟨hal-02880581⟩

Share

Metrics

Record views

44

Files downloads

41