arXiv：cs.AI· Hermann Kumbong, Ishan S. Khare, Iddah Mlauzi, Andrew Woen, Charles Ding, E. Kelly Buchanan, Simran Arora, Azalia Mirhoseini, Christopher R\'e·· 17 小时前
Next-Concept Modeling：快速并行的由粗到细文本生成
Next-Concept Modeling: Fast, Parallel Coarse-to-Fine Text Generation
AI 导读
研究者提出 Next-Concept Modeling（NCM），一种两阶段由粗到细的生成框架：多尺度残差 VQ-VAE 将文本编码为离散码层级，transformer 逐层并行预测该层全部码。
正文
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 [cs.CL] |
| (or arXiv:2610.14162v1 [cs.CL] for this version) | |
| https://doi.org/10.48550/arXiv.2610.14162 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Hermann Kumbong [view email]
[v1]
Fri, 9 Oct 2026 17:49:45 UTC (352 KB)
来源：arXiv：cs.AI · arxiv.org