PulseAugur
中
实时 09:02:56
English(EN) AgSpec: Pushing the Limits of Retrieval-Based Speculative Decoding in Coding Agent Pipelines

新方法提高 LLM 的推测解码效率 · 已追踪 10 个来源

多篇研究论文介绍了增强大型语言模型 (LLM) 推测解码的新方法,旨在提高效率和准确性。DRelay 专注于使用全局上下文修复并行草稿中的早期选择错误,而 DEdit 提出对草稿进行迭代编辑以纠正错误并扩展已接受的前缀。UBTree 结合了 unigram 和 bigram 模型来创建多样化的草稿树,而 LongSpark 开发了一种独立于前缀长度的固定成本并行草稿器。其他方法,如 SEED 重用已验证的上下文表示以加快草稿速度,HAWK 通过更好地利用视觉信息和适应目标模型变化来解决多模态草稿挑战。 AI

影响 推测解码的这些进展可以显著加快 LLM 的推理速度,使大型模型在实时应用中更加实用,并降低计算成本。

排序理由 多篇 arXiv 论文介绍了 LLM 推测解码的新研究方法。

在 Hugging Face Daily Papers 阅读 →

AI 生成摘要 · Google Gemini · 来自 10 个来源。 我们如何撰写摘要 →

新方法提高 LLM 的推测解码效率 · 已追踪 10 个来源

本文如何被排名

Signal score
0 / 100
Composite score across the factors below. Higher = stronger signal that this story matters right now.
Newsworthiness bucket
Research
多篇 arXiv 论文介绍了 LLM 推测解码的新研究方法。
Source corroboration
10 independent sources
Strong cross-source corroboration — multiple independent publishers covered this within the clustering window.
Topics
paper, infra
Editorial topic classification. Feeds into how the story surfaces on /topic/<slug> hub pages and into the per-entity coverage mix.
AI-industry relevance
High
Clearly on-topic for AI-industry coverage.
Story freshness
47 days old
Aged out of breaking-news scoring windows; ranking reflects the durable signal from the full source set.

完整方法见我们的编辑标准。

报道来源 [10]

  1. arXiv cs.AI TIER_1 English(EN) · Zhuoyu Wang, Junnan Huang, Xinyu Chen ·

    DRelay:用于前缀感知并行推测解码修复的全局草稿上下文

    arXiv:2610.01439v1 Announce Type: new Abstract: Parallel drafting reduces the drafting overhead of speculative decoding for large language models (LLMs), but its gains remain limited by the accepted prefix length. Even when the correct token is present in the candidate pool, a si…

  2. arXiv cs.CL TIER_1 English(EN) · Longxuan Yu, Bingsen Chen, Peng Shi, Dongkyu Lee, Yi Xiang, Hideo Kobayashi, Sheng Zhang, Shuaichen Chang, Xing Niu, Zhuoyan Xu, Greg Ver Steeg, Jiarong Jiang ·

    DEdit:用于推测性解码的迭代草稿编辑

    arXiv:2609.38510v1 Announce Type: new Abstract: Speculative decoding accelerates autoregressive LLMs by having a lightweight drafter propose tokens that the target model verifies in parallel. Diffusion-based drafters further reduce drafting latency by proposing multiple tokens at…

  3. arXiv cs.CL TIER_1 English(EN) · Chumeng Liang, Linxuan Wang, Xinyu Peng, Huabin Liu, Yuxin Chen, Ge Liu, Guang Lin, Qifan Song, Jianguo Li ·

    UBTree:通过单字和双字模型进行并行树草拟以实现推测性解码

    arXiv:2609.39972v1 Announce Type: new Abstract: Speculative decoding accelerates language model inference by verifying multiple draft tokens in a single target-model pass. Recent parallel drafters have achieved breakthrough performance in frontier production models, but their eff…

  4. arXiv cs.CL TIER_1 English(EN) · Jungseob Lee, Chanjun Park, Sugyeong Eo, Hyeonseok Moon ·

    为投机解码恢复离策略监督

    arXiv:2609.38795v1 Announce Type: new Abstract: Block drafters for speculative decoding are commonly trained on corpora written by external models, where a single off-policy token invalidates supervision for all subsequent slots in a block. Existing approaches discard these diver…

  5. Hugging Face Daily Papers TIER_1 English(EN) ·

    AgSpec:在编码代理管道中突破检索式推测解码的极限

    Retrieval-based speculative decoding (SD) drafts tokens by copying continuations from existing text, which suits coding agents that repeatedly reproduce code, logs, and earlier attempts. Yet existing methods fall short in agent pipelines: much of the reusable text is missing from…

  6. arXiv cs.CL TIER_1 English(EN) · Hankun Lin, Patrick Pynadath, Ruqi Zhang ·

    SEED:通过隐式编码器-解码器进行自我推测解码

    arXiv:2609.36590v1 Announce Type: new Abstract: Self-speculative decoding accelerates large language model (LLM) inference by drafting tokens from the target model itself, but faces a sharp tradeoff between the quality and cost of the draft. Early-exit methods produce drafts chea…

  7. arXiv cs.AI TIER_1 English(EN) · Rongjian Chen, Minxian Xu, Zhengxin Fang, Kejiang Ye, Chengzhong Xu ·

    DScale:通过自适应验证扩展块扩散投机解码

    arXiv:2609.37532v1 Announce Type: cross Abstract: Growing large language model applications demand efficient inference. At high concurrency, block-diffusion speculative decoding suffers from verification padding, rejected candidates, and incompatibility between variable prefixes …

  8. arXiv cs.AI TIER_1 English(EN) · Hao-Yuan He, Peng-Fei Liu, Si Shen, Ming Li ·

    LongSpark:具有固定成本并行草稿器的有效推测解码

    arXiv:2609.37029v1 Announce Type: cross Abstract: Speculative decoding accelerates autoregressive inference by verifying multiple draft tokens in a single target forward pass. However, as the context grows, existing state-of-the-art drafters become increasingly expensive, eroding…

  9. arXiv cs.AI TIER_1 English(EN) · Haohui Zhang, Keyu Chen, Haocheng Sun, Weibo Gu, Ruizhi Qiao, Xing Sun, Bo Jiang ·

    并行草稿,深度条件:用于推测解码的相邻因果注入

    arXiv:2609.36173v1 Announce Type: cross Abstract: Parallel speculative drafting generates multiple candidates in one backbone pass, but independent token selection can produce inconsistent continuations that shorten the accepted prefix. Existing methods mostly leave conditional d…

  10. arXiv cs.CV TIER_1 English(EN) · Wenhan Yang, Anirudh Rao, Ashwin Chandra ·

    HAWK:重新思考用于推测性解码的多模态草稿

    arXiv:2610.00623v1 Announce Type: new Abstract: Speculative decoding has achieved substantial lossless speedups for LLMs, but remains less effective for large vision-language models (LVLMs), where lightweight drafters struggle to use rich multimodal information. A second limitati…