Next-Concept Modeling: 高速・並列な粗視的-微視的テキスト生成
Next-Concept Modeling: Fast, Parallel Coarse-to-Fine Text Generation (arxiv.org)
要約
本論文では、テキスト生成を高速化する新しいフレームワーク「Next-Concept Modeling (NCM)」を提案しています。NCMは、テキストを階層的な離散コードにエンコードし、それを並列に生成する二段階の粗視的-微視的アプローチを採用しています。この手法は、既存の拡散言語モデル（DLM）と比較して大幅な高速化を実現しつつ、生成品質や推論タスクにおいても優れた性能を示します。
全文翻訳
Computer Science > Computation and Language
arXiv:2610.14162 (cs)
[Submitted on 9 Oct 2026]
Title:Next-Concept Modeling: Fast, Parallel Coarse-to-Fine Text Generation
Authors:[Hermann Kumbong](https://arxiv.org/search/cs?searchtype=author&query=Kumbong,+H), [Ishan S. Khare](https://arxiv.org/search/cs?searchtype=author&query=Khare,+I+S), [Iddah Mlauzi](https://arxiv.org/search/cs?searchtype=author&query=Mlauzi,+I), [Andrew Woen](https://arxiv.org/search/cs?searchtype=author&query=Woen,+A), [Charles Ding](https://arxiv.org/search/cs?searchtype=author&query=Ding,+C), [E. Kelly Buchanan](https://arxiv.org/search/cs?searchtype=author&query=Buchanan,+E+K), [Simran Arora](https://arxiv.org/search/cs?searchtype=author&query=Arora,+S), [Azalia Mirhoseini](https://arxiv.org/search/cs?searchtype=author&query=Mirhoseini,+A), [Christopher Ré](https://arxiv.org/search/cs?searchtype=author&query=R%C3%A9,+C)
View a PDF of the paper titled Next-Concept Modeling: Fast, Parallel Coarse-to-Fine Text Generation, by Hermann Kumbong and 8 other authors
[View PDF](https://arxiv.org/pdf/2610.14162) [HTML (experimental)]
Abstract:Diffusion language models (DLMs) offer parallel sampling for faster generation than autoregressive models, an increasingly important advantage as inference becomes a dominant computational workload. However, DLMs degrade quickly with fewer sampling steps, process the full sequence at every step, and require many passes over the dataset to reach good quality. We introduce Next-Concept Modeling (NCM), a two-stage, coarse-to-fine framework: a multiscale residual VQ-VAE encodes text into a hierarchy of discrete codes, and a transformer generates one level at a time, predicting all codes within a level in parallel. At matched parameter scales, NCM achieves better generative perplexity than the strongest DLM baseline on LM1B while running $21\times$ faster. On OpenWebText, NCM is competitive with the strongest DLM while running $10\times$ faster; on XSum, it matches the strongest DLM's ROUGE-L while being $31\times$ faster. On the reasoning task of Sudoku, NCM outperforms autoregressive and diffusion baselines across all difficulty levels, using fewer function evaluations. On the reasoning task of Sudoku, NCM outperforms autoregressive and diffusion baselines across all difficulty levels, using fewer function evaluations.
| Subjects: | Computation and Language (cs.CL); Artificial Intelligence (cs.AI); Machine Learning (cs.LG) |
| Cite as: | [[arXiv:2610.14162](https://arxiv.org/abs/2610.14162) [cs.CL] |
| (or [[arXiv:2610.14162v1](https://arxiv.org/abs/2610.14162v1) [cs.CL] for this version) | |
| [[https://doi.org/10.48550/arXiv.2610.14162](https://doi.org/10.48550/arXiv.2610.14162) arXiv-issued DOI via DataCite |
Submission history
From: Hermann Kumbong [[view email](https://arxiv.org/show-email/392caddc/2610.14162)]
[v1] Fri, 9 Oct 2026 17:49:45 UTC (352 KB)
Full-text links:
Access Paper:
View a PDF of the paper titled Next-Concept Modeling: Fast, Parallel Coarse-to-Fine Text Generation, by Hermann Kumbong and 8 other authors
- [[View PDF](https://arxiv.org/pdf/2610.14162)
- [[HTML (experimental)](https://arxiv.org/html/2610.14162v1)
- [[TeX Source](https://arxiv.org/src/2610.14162)
[view license](http://arxiv.org/licenses/nonexclusive-distrib/1.0/)
Additional Features
- [[Audio Summary](https://arxiv.org/audio/2610.14162)
Current browse context:
cs.CL
[< prev](https://arxiv.org/prevnext?id=2610.14162&function=prev&context=cs.CL) | [[next >](https://arxiv.org/prevnext?id=2610.14162&function=next&context=cs.CL)
[new](https://arxiv.org/list/cs.CL/new) | [[recent](https://arxiv.org/list/cs.CL/recent) | [[2026-10](https://arxiv.org/list/cs.CL/2026-10)
Change to browse by:
[cs](https://arxiv.org/abs/2610.14162?context=cs)
[cs.AI](https://arxiv.org/abs/2610.14162?context=cs.AI)
[cs.LG](https://arxiv.org/abs/2610.14162?context=cs.LG)
References & Citations
- [[SciX](https://scixplorer.org/abs/arXiv:2610.14162)
- [[Google Scholar](https://scholar.google.com/scholar_lookup?arxiv_id=2610.14162)
- [[Semantic Scholar](https://api.semanticscholar.org/arXiv:2610.14162)
Loading...
BibTeX formatted citation
loading...
Data provided by:
Bookmark
[[画像: BibSonomy](http://www.bibsonomy.org/BibtexHandler?requTask=upload&url=https://arxiv.org/abs/2610.14162&description=Next-Concept%20Modeling:%20Fast,%20Parallel%20Coarse-to-Fine%20Text%20Generation)[[画像: Reddit](https://reddit.com/submit?url=https://arxiv.org/abs/2610.14162&title=Next-Concept%20Modeling:%20Fast,%20Parallel%20Coarse-to-Fine%20Text%20Generation)
Bibliographic and Citation Tools
Bibliographic Explorer Toggle
Bibliographic Explorer ([What is the Explorer?](https://info.arxiv.org/labs/showcase.html#arxiv-bibliographic-explorer))
Connected Papers Toggle
Connected Papers ([What is Connected Papers?](https://www.connectedpapers.com/about))
Litmaps Toggle
Litmaps ([What is Litmaps?](https://www.litmaps.co/))
scite.ai Toggle
scite Smart Citations ([What are Smart Citations?](https://www.scite.ai/))
Code, Data and Media Associated with this Article
alphaXiv Toggle
alphaXiv ([What is alphaXiv?](https://alphaxiv.org/))
Links to Code Toggle
CatalyzeX Code Finder for Papers ([What is CatalyzeX?](https://www.catalyzex.com/))
DagsHub Toggle
DagsHub ([What is DagsHub?](https://dagshub.com/)
GotitPub Toggle
Gotit.pub ([What is GotitPub?](http://gotit.pub/faq))
Huggingface Toggle
Hugging Face ([What is Huggingface?](https://huggingface.co/huggingface))
ScienceCast Toggle
ScienceCast ([What is ScienceCast?](https://sciencecast.org/welcome))
Demos
Replicate Toggle
Replicate ([What is Replicate?](https://replicate.com/docs/arxiv/about))
Spaces Toggle
Hugging Face Spaces ([What is Spaces?](https://huggingface.co/docs/hub/spaces))
Spaces Toggle
TXYZ.AI ([What is TXYZ.AI?](https://txyz.ai/))
Related Papers
Recommenders and Search Tools
Link to Influence Flower
Influence Flower ([What are Influence Flowers?](https://influencemap.cmlab.dev/))
Core recommender toggle
CORE Recommender ([What is CORE?](https://core.ac.uk/services/recommender))
- 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?](https://arxiv.org/auth/show-endorsers/2610.14162) | Disable MathJax ([[What is MathJax?](https://info.arxiv.org/help/mathjax.html))