PulseAugur
实时 07:16:07
English(EN) A Symbolic Neural CPU for Quantization-Simulated Writeback and Interpretable Program Execution

新型符号神经CPU实现可解释的AI执行

研究人员开发了一种用于可解释程序执行和量化模拟写回的符号神经CPU。该架构结合了循环控制、显式操作路由器和寄存器写回,允许在执行过程中进行可见的状态转换。该系统在基准测试上精确重现了参考执行,并且即使在八位量化下也能保持符号操作路径,表明数值漂移源于连续与低精度语义不匹配,而非执行失败。进一步的增强包括ValueMemory、自适应控制器和RV32I语义桥,为可控和可解释的神经执行建立了一个框架。 AI

影响 为更透明和可控的神经执行建立了一个框架,可能提高对AI系统的信任度。

排序理由 详细介绍新型AI架构的学术论文。[lever_c_demoted from research: ic=1 ai=1.0]

在 arXiv cs.NE (Neural & Evolutionary) 阅读 →

AI 生成摘要 · Google Gemini · 来自 2 个来源。 我们如何撰写摘要 →

新型符号神经CPU实现可解释的AI执行

报道来源 [2]

  1. arXiv cs.AI TIER_1 English(EN) · Jose Luis Lima de Jesus Silva ·

    用于量化模拟写回和可解释程序执行的符号神经CPU

    arXiv:2607.10021v1 Announce Type: new Abstract: Neural networks can learn algorithmic input-output mappings, but trusting a learned executor requires more than a correct final answer because the state transitions that produce it are usually hidden. To make those transitions visib…

  2. arXiv cs.NE (Neural & Evolutionary) TIER_1 English(EN) · Jose Luis Lima de Jesus Silva ·

    A Symbolic Neural CPU for Quantization-Simulated Writeback and Interpretable Program Execution

    Neural networks can learn algorithmic input-output mappings, but trusting a learned executor requires more than a correct final answer because the state transitions that produce it are usually hidden. To make those transitions visible, we introduce a trace-supervised symbolic neu…