Model Releases
Compressing Streaming Neural Audio Encoders via Latent-Space Distillation
System-wide Dictation on Apple devices runs entirely on-device, and the speech it transcribes reaches the foundation model through a tokenizer: an encoder that maps short windows of waveform onto the
System-wide Dictation on Apple devices runs entirely on-device, and the speech it transcribes reaches the foundation model through a tokenizer: an encoder that maps short windows of waveform onto the representation the language model reads. Because that model is sparsely activated under Instruction-Following Pruning, only a small subset of its experts occupies DRAM at any time, so the always-on tokenizer competes for the same memory, and its parameter count bears directly on power and latency. In this work we study how to compress such a tokenizer by distillation, taking as the supervision…
Related
- Echo: A Joint-Embedding Predictive Architecture for Speaker Diarization and Speech Recognition in a Shared Latent Space
- Correlation-Guided Encoder Selection for Multi-Encoder Large Audio-Language Models
- Microsoft AI debuts MAI-Transcribe-2, a speech recognition model that it says beats Gemini 3.5 Transcribe and GPT-Transcribe, at $0.10/audio hour through 2026 (Michael Nuñez/VentureBeat)
- DuoMem: Towards Capable On-Device Memory Agents via Dual-Space Distillation
Source: Apple ML Research | 2026-09-24