大模型
arxiv.org ·
面向大型语言模型高效潜在空间推理的动态语义压缩
该 arXiv 论文提出 Dynamic Semantic Extraction and Inference（DSEI）框架，通过两阶段训练实现片段级隐空间推理：先以自监督学习构建 Dynamic Semantic Autoencoder（DSAE），用自适应语义加权和门控融合将片段语义压缩为紧凑隐表示，再将其集成进 LLM 架构训练模型在稠密隐空间推理。作者称在 Wanjuan 数据集上，DSEI 相比静态句级隐推理基线困惑度降低 48%，相比 token 级推理的标准 LLM 加速 2.5 倍、内存开销降低 90%。
AI推荐理由 arXiv 新提交论文（2026-09-14），提出 DSEI 框架：先用自监督学习构建 Dynamic Semantic Autoencoder（DSAE），通过自适应语义加权与门控融合把片段级语义压缩为紧凑隐表示，再将其接入 LLM 架构训练模型在稠密隐空间推理。摘要给出可核验的量化结果：在 Wanjuan 数据集上困惑度较静态句级隐推理基线降低 48%，相比 token 级推理加速 2.5 倍、内存开销降低 90%。新信息在于把隐空间推理从静态句级推进到动态片段级压缩，对 LLM 推理效率方向有参考价值；局限是仅摘要级证据、单一数据集、未见代码与第三方复现，加速与内存数字需独立验证。
72
正文
Computer Science > Computation and Language
arXiv:2609.15338 (cs)
[Submitted on 14 Sep 2026]
Title:Dynamic Semantic Compression for Efficient Latent-Space Inference in Large Language Models
Authors:Peipei Li, Dongsen Zhang, Yuchen Liu, Wenjun Xu
View a PDF of the paper titled Dynamic Semantic Compression for Efficient Latent-Space Inference in Large Language Models, by Peipei Li and 3 other authors
View PDF
HTML (experimental)
Abstract:Large Language Models (LLMs) primarily perform inference at the token level, resulting in substantial memory overhead and compromised computational efficiency. In this paper, we propose a Dynamic Semantic Extraction and Inference (DSEI) framework, which achieves segment-level inference within the latent space through a two-stage training strategy. First, we construct a Dynamic Semantic Autoencoder (DSAE) via self-supervised learning. DSAE dynamically extracts segment-level semantics and compresses them into compact latent representations via adaptive semantic weighting and gated fusion. Subsequently, we integrate the DSAE into the LLM architecture and train the model to infer over dense latent space. DSEI substantially reduces both input and generation sequences and significantly enhances inference efficiency. Extensive experiments conducted on the Wanjuan dataset demonstrate that DSEI reduces perplexity by 48% compared to static sentence-level latent inference baseline. Furthermore, compared to standard LLMs using token-level inference, DSEI accelerates inference speed by 2.5$\times$ and reduces memory overhead by 90%.
Subjects:
Computation and Language (cs.CL); Artificial Intelligence (cs.AI)
Cite as:
arXiv:2609.15338 [cs.CL]
(or
arXiv:2609.15338v1 [cs.CL] for this version)
https://doi.org/10.48550/arXiv.2609.15338
Focus to learn more
arXiv-issued DOI via DataCite (pending registration)
Submission history
From: DongSen Zhang [view email]
[v1]
Mon, 14 Sep 2026 10:24:16 UTC (2,411 KB)
Full-text links:
Access Paper:
View a PDF of the paper titled Dynamic Semantic Compression for Efficient Latent-Space Inference in Large Language Models, by Peipei Li and 3 other authors
View PDF
HTML (experimental)
TeX Source
view license
Current browse context:
cs.CL
new
|
recent
| 2026-09
Change to browse by:
cs
cs.AI
References & Citations
NASA ADS
Google Scholar
Semantic Scholar
export BibTeX citation
Loading...
BibTeX formatted citation
×
loading...
Data provided by:
Bookmark
Bibliographic Tools
Bibliographic and Citation Tools
Bibliographic Explorer Toggle
Bibliographic Explorer (What is the Explorer?)
Connected Papers Toggle
Connected Papers (What is Connected Papers?)
Litmaps Toggle
Litmaps (What is Litmaps?)
scite.ai Toggle
scite Smart Citations (What are Smart Citations?)
Code, Data, Media
Code, Data and Media Associated with this Article
alphaXiv Toggle
alphaXiv (What is alphaXiv?)
Links to Code Toggle
CatalyzeX Code Finder for Papers (What is CatalyzeX?)
DagsHub Toggle
DagsHub (What is DagsHub?)
GotitPub Toggle
Gotit.pub (What is GotitPub?)
Huggingface Toggle
Hugging Face (What is Huggingface?)
ScienceCast Toggle
ScienceCast (What is ScienceCast?)
Demos
Demos
Replicate Toggle
Replicate (What is Replicate?)
Spaces Toggle
Hugging Face Spaces (What is Spaces?)
Spaces Toggle
TXYZ.AI (What is TXYZ.AI?)
Related Papers
Recommenders and Search Tools
Link to Influence Flower
Influence Flower (What are Influence Flowers?)
Core recommender toggle
CORE Recommender (What is CORE?)
Author
Venue
Institution
Topic
About arXivLabs
arXivLabs: experimental projects with community collaborators
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
Which authors of this paper are endorsers? |
Disable MathJax (What is MathJax?)