arXivDaily arXiv每日学术速递 周一至周五更新
北美退伍军人节(Veterans Day),12 日无更新
arXiv 2602.03527cs.LG
收藏

WARP逻辑神经网络

WARP Logic Neural Networks

Princeton University(普林斯顿大学)

Lino Gerlach, Thore Gerlach, Liv Våge, Elliott Kauffman, Isobel Ojalvo

AI总结:

WARP逻辑神经网络通过引入可学习阈值和残差初始化,实现了高效且精确的布尔函数学习,提升了训练效率和模型扩展能力。

摘要:

快速且高效的AI推理正变得越来越重要,而近期直接学习低级逻辑操作的模型已取得了最先进的性能。然而,现有的逻辑神经网络存在训练成本高、引入冗余或依赖近似梯度的问题,限制了可扩展性。为克服这些限制,我们引入了WAlsh松弛用于概率(WARP)逻辑神经网络——一种新的基于梯度的框架,能够高效地学习硬件原生逻辑块的组合。我们证明WARP为精确学习布尔函数提供了最参数高效的表示,并且几种先前的方法作为受限的特殊情形出现。通过引入可学习的阈值和残差初始化,训练得到了改进,同时通过随机平滑弥合了松弛训练与离散逻辑推理之间的差距。实验表明,比最先进的基线方法收敛更快,同时能够有效扩展到更深的架构和更高输入 arity 的逻辑函数。

英文摘要:

Fast and efficient AI inference is increasingly important, and recent models that directly learn low-level logic operations have achieved state-of-the-art performance. However, existing logic neural networks incur high training costs, introduce redundancy or rely on approximate gradients, which limits scalability. To overcome these limitations, we introduce WAlsh Relaxation for Probabilistic (WARP) logic neural networks -- a novel gradient-based framework that efficiently learns combinations of hardware-native logic blocks. We show that WARP yields the most parameter-efficient representation for exactly learning Boolean functions and that several prior approaches arise as restricted special cases. Training is improved by introducing learnable thresholding and residual initialization, while we bridge the gap between relaxed training and discrete logic inference through stochastic smoothing. Experiments demonstrate faster convergence than state-of-the-art baselines, while scaling effectively to deeper architectures and logic functions with higher input arity.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑