REVIEW 3 cited by
Neural Speech Extraction with Human Feedback
Not yet reviewed by Pith; the record is open.
This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.
SPECIMEN: schema-true, not a live event
T0 review · schema-true
One-sentence machine reading of the paper's core claim.
pith:XXXXXXXX · record.json · timestamp
Neural Speech Extraction with Human Feedback
read the original abstract
We present the first neural target speech extraction (TSE) system that uses human feedback for iterative refinement. Our approach allows users to mark specific segments of the TSE output, generating an edit mask. The refinement system then improves the marked sections while preserving unmarked regions. Since large-scale datasets of human-marked errors are difficult to collect, we generate synthetic datasets using various automated masking functions and train models on each. Evaluations show that models trained with noise power-based masking (in dBFS) and probabilistic thresholding perform best, aligning with human annotations. In a study with 22 participants, users showed a preference for refined outputs over baseline TSE. Our findings demonstrate that human-in-the-loop refinement is a promising approach for improving the performance of neural speech extraction.
Forward citations
Cited by 3 Pith papers
-
Unmixing The Crowd: Learning Persistent Speaker Representations from Mixture-Derived Multi-Speaker Embeddings
A neural model predicts a set of speaker embeddings from noisy mixtures to enable enrollment-free target speech extraction, outperforming baselines on LibriMix and generalizing to real recordings.
-
Unmixing The Crowd: Learning Persistent Speaker Representations from Mixture-Derived Multi-Speaker Embeddings
A LoRA-generating hypernetwork unifies conventional pulmonary and opportunistic cardiac screening from non-contrast chest CT, matching single-task models while outperforming multi-task baselines on prospective multi-c...
-
GenTSE: Enhancing Target Speaker Extraction via a Coarse-to-Fine Generative Language Model
A two-stage decoder-only language model with continuous embeddings and UTMOS-based preference fine-tuning reports improved target-speaker-extraction scores on Libri2Mix.
discussion (0)
Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.