Research map / Probabilistic and scientific modeling / Variational and generative sampling
Joint PLDA for Simultaneous Modeling of Two Factors
Luciana Ferrer, Mitchell McLaren
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 |
|---|---|---|---|
| use JPLDA instead of FPLDA baseline when each speaker has a single language in training | PLDA speaker verification · Mixer and LASRS (SINGLE-LAN training) | relative gain over FPLDA: gains from 17% of up to 67% relative to the FPLDA baseline | Improved |
| use tied PLDA (two-component mixture) instead of JPLDA on mismatched-channel test data | TPLDA · LASRS channels | EER | Hurt |
| add smart initialization to JPLDA and run only one EM iteration (ablation) | JPLDA · Mixer test (FULL training) | EER: 1 EM iteration instead of 50 | Improved |
| set same-language prior to values between 0.2 and 0.8 (ablation) | JPLDA · Mixer test (FULL training) | EER: 0.2 to 0.8 give almost identical performance | No clear effect |
| train FPLDA on subset that discards telephone data from English-only speakers (ablation) | FPLDA · Mixer and LASRS (FULL and SINGLE-LAN) | EER | Improved |
| train JPLDA on full set instead of subset (ablation) | JPLDA · SINGLE-LAN training | EER | Improved |
| use training subset for FPLDA when measuring Cllr (ablation) | FPLDA · Mixer and LASRS | Cllr | Mixed |
Closest papers by meaning
- Uncertain LDA: Including Observation Uncertainties in Discriminative Transforms (TPAMI 2016 · similarity 0.85)
- Separability-Oriented Subclass Discriminant Analysis (TPAMI 2017 · similarity 0.84)
- Joint Discriminative Bayesian Dictionary and Classifier Learning (CVPR 2017 · similarity 0.83)
- Discriminant Feature Extraction by Generalized Difference Subspace (TPAMI 2022 · similarity 0.83)
- On Sparse Canonical Correlation Analysis (NeurIPS 2024 · similarity 0.82)
- Int*-Match: Balancing Intra-Class Compactness and Inter-Class Discrepancy for Semi-Supervised Speaker Recognition (AAAI 2025 · similarity 0.81)
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.
