paper-with-me

홈 › Papers

Scaling Properties of Continuous Diffusion Spoken Language Models

2026-04-27 · Jason Ramapuram, Eeshan Gunesh Dhekane, Amitis Shidani, Dan Busbridge, Bogdan Mazoure, Zijin Gu, Russ Webb, Tatiana Likhomanenko, Navdeep Jaitly arxiv

Speech-only spoken language models (SLMs) lag behind text and text-speech models in performance, with recent discrete autoregressive (AR) SLMs indicating significant computational and data demands to match text models. Since discretizing continuous speech for AR creates bottlenecks, we explore whether continuous diffusion (CD) SLM is more viable. To quantify the SLMs linguistic quality, we introduce the phoneme Jensen-Shannon divergence (pJSD) metric. Our analysis reveals CD SLMs, mirroring AR behavior, exhibit scaling laws for validation loss and pJSD, and show optimal token-to-parameter ratios decreasing as compute scales. However, for the latter, loss becomes insensitive to choice of data and model sizes, showing potential for fast inference. Scaling CD SLMs to 16B parameters with tens of millions of hours of conversational data enables generation of emotive, prosodic, multi-speaker, multilingual speech, though achieving long-form coherence remains a significant challenge.

📄 PDF Abstract BibTeX arXiv:2604.24416

Code (0)

등록된 구현이 없습니다.

Similar Papers 제목 키워드 기반

Towards Continuous Sign Language Conversation from Isolated Signs

2026-05-14 · Youngmin Kim, Kyobin Choo, Jiwoo Park, Minseo Kim 외 arxiv

Sign language is the primary language for many Deaf and Hard-of-Hearing (DHH) signers, yet most conversational AI systems still mediate interaction through spoken or written language. This spoken-language-centered interf…

Domain Generalization

Scaling Properties of Text Conditioning in Visual Generation

2026-07-31 · Zilong Chen, Chaorui Deng, Kunchang Li, Hongyi Yuan 외 hf

We study empirical scaling properties for text conditioning in visual generation. Such properties have rarely been measured because diffusion loss does not scale with the number of tokens in natural-language prompts. Sur…

MS2SL: Multimodal Spoken Data-Driven Continuous Sign Language Production

2024-07-04 · Jian Ma, Wenguan Wang, Yi Yang, Feng Zheng

Sign language understanding has made significant strides; however, there is still no viable solution for generating sign sequences directly from entire spoken content, e.g., text or speech. In this paper, we propose a un…

Sign Language Production

Self-conditioned Embedding Diffusion for Text Generation

2022-11-08 · Robin Strudel, Corentin Tallec, Florent Altché, Yilun Du 외

Can continuous diffusion models bring the same performance breakthrough on natural language they did for image generation? To circumvent the discrete nature of text data, we can simply project tokens in a continuous spac…

Image GenerationLanguage ModelingLanguage ModellingText Generation

Continuous Latent Diffusion Language Model

2026-05-07 · Hongcan Guo, Qinyu Zhao, Yian Zhao, Shen Nie 외 arxiv

Large language models have achieved remarkable success under the autoregressive paradigm, yet high-quality text generation need not be tied to a fixed left-to-right order. Existing alternatives still struggle to jointly …

Representation LearningText Generation