提出可收敛到词嵌入的连续语言模型，无需交叉熵训练解码器。
ConvergeFlow: Language Flow with Provable Convergence to Token Embeddings
- 在词嵌入凸包内约束预测器，用流匹配的均方误差训练。
- 理论证明即使预测有误，流仍能收敛到有效词嵌入。
- 支持多种采样策略，平衡生成困惑度与熵，适合研究者使用。
最近的连续扩散与基于流的语言模型（LMs）性能已达到与离散模型相当水平。然而，现有连续框架仍依赖交叉熵（CE）监督的解码器，因为流轨迹无法保证终止于有效词嵌入。为此，我们提出 extbf{ConvergeFlow}，一种嵌入空间的流式语言模型，将数据预测器限制在词嵌入的凸包内，并仅使用流匹配诱导的均方误差目标进行训练。在适当的正则条件下，我们证明了即便数据预测存在误差，该流仍能收敛至有效词嵌入，从而实现无需CE监督解码器的直接词预测。此外，我们设计了三种采样机制，用于调控生成困惑度与熵之间的权衡。在OpenWebText上的实验表明，ConvergeFlow性能可媲美现有的连续与离散扩散语言模型。这些结果展示了流式范式在语言建模中的潜力。代码已公开于https://github.com/Na-Li66/ConvergeFlow。
原文摘要 · Abstract (English)
Recent advances in continuous diffusion and flow-based language models (LMs) have achieved performance competitive with discrete LMs. However, existing continuous frameworks still rely on decoders supervised with cross entropy (CE) because the flow trajectories are not guaranteed to terminate at valid token embeddings. Motivated by this limitation, we introduce \textbf{ConvergeFlow}, an embedding-space flow-based LM, which constrains the data predictor to the convex hull of token embeddings and trains it solely with the mean squared error objective induced by flow matching. Under suitable regularity conditions, we prove that the resulting flow converges to valid token embeddings despite errors in the data predictor, enabling direct token prediction without a CE-supervised decoder. We further develop three sampling mechanisms for controlling the trade-off between the generative perplexity and entropy. Experiments on OpenWebText demonstrate that ConvergeFlow achieves performance competitive with existing continuous and discrete diffusion LMs. These findings demonstrate the potential of the flow-based paradigm for language modeling. Our code is available at https://github.com/Na-Li66/ConvergeFlow.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。