A Generic Model for Privacy-Preserving Authentication on Smartphones
Bibliographic record
Abstract
With the increasing use of biometrics for user authentication especially on mobile devices, its privacy and resource requirements are becoming big challenges to consider. In this paper, we propose a generic model for privacy-preserving yet accurate authentication on smartphones using the mobile matching on card (MMOC) technique and transfer learning. MMOC technique takes advantage of SIM cards as a secure element (SE) on smartphones to increase the security and privacy of user verification with low performance overhead. In order to improve the performance accuracy of the system, we use transfer learning and fine-tune a network suitable for implementation on off-the-shelf SIM cards available on smartphones. The classification sub-network is migrated to the SIM card for a lightweight and secure user verification. However, the implementation of classification sub-network on constrained resource smart cards with high accuracy and efficiency is a challenging task. We propose log quantization scheme and an on-card optimization architecture to speed-up the forward pass of the sub-network and retain the system's accuracy close to the original model with low memory footprint and real-time verification response. Using a public mobile face dataset, we evaluate our privacy-preserving verification system. Our results show that the proposed system achieves Equal Error Rate (EER) of 0.4%-2% in real-time, with response time of 1.5 seconds.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.001 | 0.002 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.000 | 0.001 |
| Science and technology studies | 0.000 | 0.001 |
| Scholarly communication | 0.001 | 0.003 |
| Open science | 0.001 | 0.002 |
| Research integrity | 0.001 | 0.001 |
| Insufficient payload (model declined to judge) | 0.003 | 0.001 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".