End-To-End Detection Of Attacks To Automatic Speaker Recognizers With Time-Attentive Light Convolutional Neural Networks
Bibliographic record
Abstract
In this contribution, we introduce convolutional neural network architectures aiming at performing end-to-end detection of attacks to voice biometrics systems, i.e. the model provides scores corresponding to the likelihood of attack given general purpose time-frequency features obtained from speech. Microphone level attackers based on speech synthesis and voice conversion techniques are considered, along with presentation replay attacks. While the convolutional models yield a sequence of representations corresponding to different parts of the input at varying time steps, concatenated first and second-order statistics pooled from the outputs of a self-attention layer are used as a fixed-dimension representations of utterances of varying length, which are then input into a set of fully connected layers to finally yield scores. Evaluation of the proposed framework is performed with data from ASVspoof 2019 challenge yielding relative improvements higher than one order of magnitude in terms of equal error rate over two baseline systems provided by ASVspoof 2019's organizers, and significant improvements over the benchmark systems we evaluated.
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 | 0.003 |
| Meta-epidemiology (narrow) | 0.002 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.000 |
| Bibliometrics | 0.001 | 0.000 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.001 | 0.001 |
| Open science | 0.001 | 0.002 |
| Research integrity | 0.001 | 0.001 |
| Insufficient payload (model declined to judge) | 0.003 | 0.003 |
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".