Skip to Main content Skip to Navigation
Conference papers

Unsupervised labelling of stolen handwritten digit embeddings with density matching

Abstract : Biometrics authentication is now widely deployed, and from that omnipresence comes the necessity to protect private data. Recent studies proved touchscreen handwritten digits to be a reliable biomet-rics. We set a threat model based on that biometrics: in the event of theft of unlabelled embeddings of handwritten digits, we propose a labelling method inspired by recent unsupervised translation algorithms. Provided a set of unlabelled embeddings known to have been produced by a Long Short Term Memory Recurrent Neural Network (LSTM RNN), we demonstrate that inferring their labels is possible. The proposed approach involves label-wise clustering of the embeddings and label identification of each group by matching their distribution to the label-relative classes of a comparison hand-crafted labeled set of embeddings. Cluster labelling is done through a two steps process including a genetic algorithm that finds the N-best matching hypotheses before a fine-tuning of those N-candidates. The proposed method was able to infer the correct labels on 100 randomised runs on different dataset splits.
Complete list of metadata

Cited literature [24 references]  Display  Hide  Download
Contributor : anthony larcher Connect in order to contact the contributor
Submitted on : Wednesday, July 22, 2020 - 7:18:16 PM
Last modification on : Wednesday, January 19, 2022 - 12:00:02 PM
Long-term archiving on: : Tuesday, December 1, 2020 - 5:15:36 AM


Files produced by the author(s)


  • HAL Id : hal-02904938, version 1



Thomas Thebaud, Gaël Le Lan, Anthony Larcher. Unsupervised labelling of stolen handwritten digit embeddings with density matching. International Workshop on Security in Machine Learning and its Applications (SiMLA), Oct 2020, Rome, Italy. ⟨hal-02904938⟩



Record views


Files downloads