paper-with-me

Papers

Thinking-while-speaking: A Controlled, Interleaved Reasoning Method for Real-Time Speech Generation

2026-05-20 · Xuan Du, Qiangyu Yan, Wenshuo Li, Borui Jiang, Changming Xiao, Han Shu, Xinghao Chen arxiv

The thinking-while-speaking paradigm aims to make AI communication more human. A key challenge is maintaining fluent speech while performing deep reasoning. Our method, InterRS, tackles this by inserting reasoning steps only during natural speech generation. This requires high-quality data where reasoning and speech are precisely aligned, and the length ratio are under controlled. We introduce a novel pipeline to generate such seamlessly interleaved audio data. To train our model, we combine interleaved SFT with refined data and reinforcement learning with two new rewards: a TA-Balance Reward to manage timing and thinking-answer ratio, and a Linguistic Quality Reward to refine expression. Experiments show our approach achieves 13% better performance on mathmatical and logic benchmarks while generating instant response like a spoken-language instruct model which outputs fast CoT response. Furthermore, our method generates more natural and fluent answers than prior methods.

📄 PDF Abstract BibTeX arXiv:2605.20946

Code (0)

등록된 구현이 없습니다.

Tasks

Reinforcement Learning

Similar Papers 제목 키워드 기반

Mini-Omni-Reasoner: Token-Level Thinking-in-Speaking in Large Speech Models

2025-08-18 · Zhifei Xie, Ziyang Ma, Zihang Liu, Kaiyu Pang 외 arxiv

Reasoning is essential for effective communication and decision-making. While recent advances in LLMs and MLLMs have shown that incorporating explicit reasoning significantly improves understanding and generalization, re…

Arithmetic Reasoning

DeepTool: Scaling Interleaved Deliberation in Tool-Integrated Reasoning via Process-Supervised Reinforcement Learning

2026-05-28 · Yang He, Xiao Ding, Bibo Cai, Yufei Zhang 외 arxiv

Tool-Integrated Reasoning (TIR) extends LLM capabilities by leveraging external environments. However, existing methods lack the deliberation during sequential tool invocation required for strategic planning and self-cor…

Reinforcement Learning

Thinking-while-Generating: Interleaving Textual Reasoning throughout Visual Generation

2025-11-20 · Ziyu Guo, Renrui Zhang, Hongyu Li, Manyuan Zhang 외 arxiv

Recent advances in visual generation have increasingly explored the integration of reasoning capabilities. They incorporate textual reasoning, i.e., think, either before (as pre-planning) or after (as post-refinement) th…

Reinforcement Learning

AdaptR1: Reinforcement Learning Based Adaptive Interleaved Thinking in Multi-hop Question Answering

2026-05-29 · Yuxin Wang, Jiahao Lu, Qifeng Wu, Shicheng Fang 외 arxiv

Large Language Models (LLMs) have achieved remarkable performance in complex reasoning tasks through Chain-of-Thought (CoT) prompting. However, this approach often leads to ``over-thinking,'' where models generate unnece…

Multi-hop Question AnsweringReinforcement Learning

Reliable Thinking with Images

2026-02-13 · Haobin Li, Yutong Yang, Yijie Lin, Xiang Dai 외 arxiv

As a multimodal extension of Chain-of-Thought (CoT), Thinking with Images (TWI) has recently emerged as a promising avenue to enhance the reasoning capability of Multi-modal Large Language Models (MLLMs), which generates…