# ELEAT-SAGA: Early & Late Integration with Evading Alternating Training for Spoof-Robust Speaker Verification

arXiv: https://arxiv.org/abs/2602.13761 · HTML: https://mlsanity.com/paper/2602.13761/

## Key facts

- No citing papers in the map yet (data through 19 Oct 2026).
- Placed in Anti-spoofing and adversarial audio attacks, one of 11 directions in Speech and audio language models.
- Builds on 19 of its 32 references that are in the map.
- First posted on arXiv 14 Feb 2026.

## Placed in

- Speech and audio language models / [Anti-spoofing and adversarial audio attacks](https://mlsanity.com/speech/anti-spoofing-and-adversarial-audio-attacks/)

## Builds on (most cited first, 19 of 19)

- [ECAPA-TDNN: Emphasized Channel Attention, Propagation and Aggregation in TDNN Based Speaker Verification](https://mlsanity.com/paper/2005.07143/) (2020)
- [GradNorm: Gradient Normalization for Adaptive Loss Balancing in Deep Multitask Networks](https://mlsanity.com/paper/1711.02257/) (2017)
- [AASIST: Audio Anti-Spoofing using Integrated Spectro-Temporal Graph Attention Networks](https://mlsanity.com/paper/2110.01200/) (2021)
- [End-to-End Spectro-Temporal Graph Attention Networks for Speaker Verification Anti-Spoofing and Speech Deepfake Detection](https://mlsanity.com/paper/2107.12710/) (2021)
- [SpoofCeleb: Speech Deepfake Detection and SASV In The Wild](https://mlsanity.com/paper/2409.17285/) (2024)
- [SASV 2022: The First Spoofing-Aware Speaker Verification Challenge](https://mlsanity.com/paper/2203.14732/) (2022)
- [Golden Gemini is All You Need: Finding the Sweet Spots for Speaker Verification](https://mlsanity.com/paper/2312.03620/) (2023)
- [Frequency and Multi-Scale Selective Kernel Attention for Speaker Verification](https://mlsanity.com/paper/2204.01005/) (2022)
- [a-DCF: an architecture agnostic metric with application to spoofing-robust speaker verification](https://mlsanity.com/paper/2403.01355/) (2024)
- [AASIST3: KAN-Enhanced AASIST Speech Deepfake Detection using SSL Features and Additional Regularization for the ASVspoof 2024 Challenge](https://mlsanity.com/paper/2408.17352/) (2024)
- [Baseline Systems for the First Spoofing-Aware Speaker Verification Challenge: Score and Embedding Fusion](https://mlsanity.com/paper/2204.09976/) (2022)
- [A Probabilistic Fusion Framework for Spoofing Aware Speaker Verification](https://mlsanity.com/paper/2202.05253/) (2022)
- [Towards single integrated spoofing-aware speaker verification embeddings](https://mlsanity.com/paper/2305.19051/) (2023)
- [On the potential of jointly-optimised solutions to spoofing attack detection and automatic speaker verification](https://mlsanity.com/paper/2209.00506/) (2022)
- [Generalizing Speaker Verification for Spoof Awareness in the Embedding Space](https://mlsanity.com/paper/2401.11156/) (2024)
- [To what extent can ASV systems naturally defend against spoofing attacks?](https://mlsanity.com/paper/2406.05339/) (2024)
- [The Vicomtech Spoofing-Aware Biometric System for the SASV Challenge](https://mlsanity.com/paper/2204.01399/) (2022)
- [PTEENet: Post-Trained Early-Exit Neural Networks Augmentation for Inference Cost Optimization](https://mlsanity.com/paper/2501.02508/) (2025)
- [ATMM-SAGA: Alternating Training for Multi-Module with Score-Aware Gated Attention SASV system](https://mlsanity.com/paper/2505.18273/) (2025)

## Abstract (arXiv)

Spoofing-robust automatic speaker verification (SASV) seeks to build automatic speaker verification systems that are robust against both zero-effort impostor attacks and sophisticated spoofing techniques such as voice conversion (VC) and text-to-speech (TTS). In this work, we propose a novel SASV architecture that introduces score-aware gated attention (SAGA), SASV-SAGA, enabling dynamic modulation of speaker embeddings based on countermeasure (CM) scores. By integrating speaker embeddings and CM scores from pre-trained ECAPA-TDNN and AASIST models respectively, we explore several integration strategies including early, late, and full integration. We further introduce alternating training for multi-module (ATMM) and a refined variant, evading alternating training (EAT). Experimental results on the ASVspoof 2019 Logical Access (LA) and Spoofceleb datasets demonstrate significant improvements over baselines, achieving a spoofing aware speaker verification equal error rate (SASV-EER) of 1.22% and minimum normalized agnostic detection cost function (min a-DCF) of 0.0304 on the ASVspoof 2019 evaluation set. These results confirm the effectiveness of score-aware attention mechanisms and alternating training strategies in enhancing the robustness of SASV systems.