AI 中文总结
该研究在直觉主义蕴含逻辑中建模生成式AI的下一个令牌预测，推导等价于乘法RNN的神经架构，用Prolog定理证明器验证模型性质，并将其与Transformer等模型对比。
AI 中文摘要
我们在直觉主义蕴含逻辑中对当今生成式AI的关键使能技术——自回归因果神经网络中的下一个令牌预测进行建模。在我们的框架中，下一个令牌预测对应于肯定前件式，序列处理则在柯里-霍华德对应下成为构造性证明扩展。我们基于Prolog的专用定理证明器验证了神经模型的基本性质，其中包括交换与非交换序列以及单令牌与多令牌预测选择之间的关系。我们从下一个令牌预测作为嵌套直觉主义蕴含的证明论解释中推导出一种等价于乘法循环神经网络（multiplicative RNNs）的神经架构，并将该模型与Transformer、状态空间模型及递归大语言模型（recursive LLMs）进行定位比较。
英文摘要
We model in intuitionistic implicational logic the key enabler of today's GenerativeAI: the next-token prediction in autoregressive causal neural networks. In our framework, next-token prediction corresponds to modus ponens, and sequence processing becomes constructive proof extension under the Curry-Howard correspondence. Our Prolog-based specialized theorem provers validate fundamental properties of the neural models, among which relations between commutative vs. non-commutative sequencing and single-token vs. multi-token prediction choices. We derive a neural architecture equivalent to multiplicative RNNs that arises naturally from a proof-theoretic interpretation of next-token prediction as nested intuitionistic implication and position the model relative to transformers, state-space models and recursive LLMs.
CommentsIn Proceedings ICLP 2026, arXiv:2607.17707. Dedicated to the memory of Paul Tarau - he sadly passed away about one month after having submitting this paper
Journal refEPTCS 450, 2026, pp. 52-66
