G-STAR: End-to-End Global Speaker-Tracking Attributed Recognition
This addresses the challenge of robust speaker identity linking and fine-grained temporal boundaries in multi-party speech recognition, which is incremental over prior Speech-LLM systems.
The paper tackles the problem of timestamped speaker-attributed ASR for long-form, multi-party speech with overlap, aiming to preserve meeting-level speaker identity consistency while producing accurate transcripts. The result is G-STAR, an end-to-end system that integrates a speaker-tracking module with a Speech-LLM backbone, achieving improved performance in experiments analyzing cue fusion and trade-offs.
We study timestamped speaker-attributed ASR for long-form, multi-party speech with overlap, where chunk-wise inference must preserve meeting-level speaker identity consistency while producing time-stamped, speaker-labeled transcripts. Previous Speech-LLM systems tend to prioritize either local diarization or global labeling, but often lack the ability to capture fine-grained temporal boundaries or robust cross-chunk identity linking. We propose G-STAR, an end-to-end system that couples a time-aware speaker-tracking module with a Speech-LLM transcription backbone. The tracker provides structured speaker cues with temporal grounding, and the LLM generates attributed text conditioned on these cues. G-STAR supports both component-wise optimization and joint end-to-end training, enabling flexible learning under heterogeneous supervision and domain shift. Experiments analyze cue fusion, local versus long-context trade-offs and hierarchical objectives.