PS4: Proxy-Supervised Joint Training for Real Target Speaker Extraction
For researchers working on speaker extraction in real-world conversations, PS4 provides a practical training method without requiring clean target speech, achieving competitive results on a public benchmark.
PS4 introduces a proxy-supervised training framework for target speaker extraction in real conversational mixtures, using a large-scale corpus of 71,771 samples and four differentiable objectives. It achieves 2nd place on the REAL-T challenge leaderboard with the best speaker similarity and timing F1.
Training target speaker extraction (TSE) models for real conversational mixtures remains challenging because large-scale training corpora and clean target speech for supervision are unavailable. We present PS4, a proxy-supervised training framework for TSE in real conversational mixtures, with two main contributions. First, we construct a large-scale corpus of 71,771 training samples derived from four public datasets, covering both Chinese and English scenarios. Each sample contains an overlapping speech mixture, per-speaker enrollment audio, a ground-truth transcript, and frame-level voice activity labels. Second, we propose a proxy-supervised joint training strategy that fine-tunes a BSRNN-based TSE model using four complementary differentiable objectives: ASR cross-entropy, speaker similarity, frame-level voice activity detection, and perceptual audio quality. Starting from a publicly available pre-trained checkpoint, only the BSRNN separator is updated during fine-tuning. On the REAL-T challenge leaderboard, PS4 ranks 2nd overall, achieving the best speaker similarity and timing F1 among all submitted systems.