Research map / Label-efficient and robust learning / Recognition and long-tail
Beyond Speech: Generalizing D-Vectors for Biometric Verification
Jacob Baldwin, Ryan Burnham, Andrew Meyer, Robert Dora, Robert Wright
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| combine cosine, L2 and z-score similarities instead of cosine only (ablation) | D-Vector · MultiMod keystroke | EER %: 11.8 -> 11.5 | Improved |
| combine three similarities instead of cosine only (ablation) | D-Vector · Clarkson keystroke | EER %: 16.1 -> 15.3 | Improved |
| combine three similarities instead of cosine only (ablation) | D-Vector · MultiMod gait | EER %: 18.0 -> 17.5 | Improved |
| combine three similarities instead of cosine only (ablation) | D-Vector · UCI gait | EER %: 19.3 -> 15.3 | Improved |
| use L2 only instead of cosine on UCI (ablation) | D-Vector · UCI gait | EER %: 19.3 -> 15.0 | Improved |
| use L2 only instead of cosine (ablation) | D-Vector · MultiMod gait | EER %: 18.0 -> 30.8 | Hurt |
| use five test samples instead of one (ablation) | D-Vector · MultiMod gait | EER %: 17.5 -> 7.0 | Improved |
| use five test samples instead of one | D-Vector · UCI gait | EER %: 15.3 -> 11.4 | Improved |
| use five test samples instead of one | D-Vector · Clarkson keystroke | EER %: 15.3 -> 8.7 | Improved |
| use five test samples instead of one | Intel gait baseline · MultiMod gait | EER %: 27.0 -> 24.2 | No clear effect |
| replace G&P with D-Vector | keystroke verification · MultiMod | EER %: 26.3 -> 7.7 | Improved |
| replace Intel GMM with D-Vector DNN | gait verification · UCI | EER %: 28.2 -> 15.3 | Improved |
Closest papers by meaning
- Certification of Speaker Recognition Models to Additive Perturbations (AAAI 2025 · similarity 0.83)
- Improving Generalization for AI-Synthesized Voice Detection (AAAI 2025 · similarity 0.81)
- Test-Time Domain Generalization for Face Anti-Spoofing (CVPR 2024 · similarity 0.80)
- Unmasking Puppeteers: Leveraging Biometric Leakage to Expose Impersonation in AI-Based Videoconferencing (NeurIPS 2025 · similarity 0.80)
- SpeechForensics: Audio-Visual Speech Representation Learning for Face Forgery Detection (NeurIPS 2024 · similarity 0.79)
- MMJ-Bench: A Comprehensive Study on Jailbreak Attacks and Defenses for Vision Language Models (AAAI 2025 · similarity 0.79)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
