arXiv Artificial Intelligence

Acoustic-to-Text KV Compression for Full-Duplex Speech Models

2 hours ago 28.09.2026 07:00 👁 2 views
Acoustic-to-Text KV Compression for Full-Duplex Speech Models

Quick summary

arXiv:2609.31224v1 Announce Type: cross Abstract: Full-duplex speech language models continuously accumulate acoustic key-value (KV) states, making long-running interactions memory-intensive. During listening, the model can finish processing an audio unit before the next arrives; we term the remaining interval listening-time slack. We propose acoustic-to-text KV compression, which introduces a transcription side channel to convert incoming speech into compact textual memory within this interval. When the cache exceeds a target budget during inference, older acoustic states are evicted while tr

Key takeaways

  • arXiv:2609.31224v1 Announce Type: cross Abstract: Full-duplex speech language models continuously accumulate acoustic key-value (KV) states, making long-running interactions memory-intensive.
  • During listening, the model can finish processing an audio unit before the next arrives; we term the remaining interval listening-time slack.
  • We propose acoustic-to-text KV compression, which introduces a transcription side channel to convert incoming speech into compact textual memory within this interval.

Why it matters

“Acoustic-to-Text KV Compression for Full-Duplex Speech Models” should be evaluated beyond branding and benchmark scores. Its practical importance will emerge in task accuracy, latency, unit cost, safety and integration with real workflows.

Kaynak sitede devamını oku: arXiv Artificial Intelligence ↗