8 papers
Scaling Properties of Continuous Diffusion Spoken Language Models
Jason Ramapuram, Eeshan Gunesh Dhekane, Amitis Shidani +6
Speech-only spoken language models (SLMs) lag behind text and text-speech models in performance, with recent discrete autoregressive (AR) SLMs indicating significant computational…
Path-Constrained Mixture-of-Experts
Zijin Gu, Tatiana Likhomanenko, Vimal Thilak +2
Sparse Mixture-of-Experts (MoE) architectures route each token through a subset of experts at each layer independently. We propose viewing MoE computation through the lens of \emph…
Which Data Matter? Embedding-Based Data Selection for Speech Recognition
Zakaria Aldeneh, Skyler Seto, Maureen de Seyssel +8
Modern ASR systems are typically trained on large-scale pseudo-labeled, in-the-wild data spanning multiple domains. While such heterogeneous data benefit generalist models designed…
Closing the Gap Between Text and Speech Understanding in LLMs
Santiago Cuervo, Skyler Seto, Maureen de Seyssel +5
Large Language Models (LLMs) can be adapted to extend their text capabilities to speech inputs. However, these speech-adapted LLMs consistently underperform their text-based counte…
ChipChat: Low-Latency Cascaded Conversational Agent in MLX
Tatiana Likhomanenko, Richard He Bai, Zijin Gu +8
The emergence of large language models (LLMs) has transformed spoken dialog systems, yet the optimal architecture for real-time on-device voice agents remains an open question. Whi…
Omni-Router: Sharing Routing Decisions in Sparse Mixture-of-Experts for Speech Recognition
Zijin Gu, Tatiana Likhomanenko, Navdeep Jaitly
Mixture-of-experts (MoE) architectures have expanded from language modeling to automatic speech recognition (ASR). Traditional MoE methods, such as the Switch Transformer, route ex…
SpeakStream: Streaming Text-to-Speech with Interleaved Data
Richard He Bai, Zijin Gu, Tatiana Likhomanenko +1
The latency bottleneck of traditional text-to-speech (TTS) systems fundamentally hinders the potential of streaming large language models (LLMs) in conversational AI. These TTS sys…
What Makes the Preferred Thinking Direction for LLMs in Multiple-choice Questions?
Yizhe Zhang, Richard Bai, Zijin Gu +5
Language models usually use left-to-right (L2R) autoregressive factorization. However, L2R factorization may not always be the best inductive bias. Therefore, we investigate whethe…