发表机构
Newcastle University; Northumbria University; University of Agder（纽卡斯尔大学; 诺森比亚大学; 阿格德尔大学）
机构由 AI 辅助整理，请以论文原文为准。AI 中文总结
提出CRISP框架，将二元神经网络的最后一层激活向量重构为Tsetlin Machine子句，提供可审计的符号决策轨迹，实验验证其保真度与可解释性。
AI 中文摘要
深度神经网络通过分层数值变换实现高精度，但其决策难以审计，因为决策证据编码在隐藏激活中而非显式规则中。本文提出CRISP框架，将二元神经教师网络的最后一层激活向量（LLAV）重构为Tsetlin Machine（TM）子句。CRISP对教师网络倒数第二层预逻辑激活进行符号二值化，并为每个LLAV神经元分配一个独立Tsetlin Machine（ITM）。每个重构的隐藏位由布尔化输入特征上的命题子句表示，从而提供从命名输入阈值到命名教师神经元的直接符号轨迹。CRISP在MNIST、KMNIST、FashionMNIST（FMNIST）、SVHN和CIFAR10上进行了评估，使用BinaryConnect卷积神经网络（BCCNN）教师和全二元神经网络（BNN）教师，并额外研究了多种位深度下的二元阈值化、温度计编码和四分位分箱。结果表明，在测试的BNN设置中，LLAV符号二值化不会降低教师头部精度，而ITM重构误差是主要限制因素。四分位单比特布尔化在SVHN上实现了最强的重构保真度，测试保真度为87.52%，在CIFAR10上具有竞争力，重构的LLAV在FMNIST上保留了78.41%的教师头部精度。合并子句证据可视化显示，学习的ITM文字集中在中心化基准中的目标区域。因此，CRISP为检查二元神经教师网络的最终隐藏表示提供了一条子句级路径。
英文摘要
Deep neural networks achieve high accuracy through layered numerical transformations, yet their decisions remain difficult to audit because decision evidence is encoded in hidden activations rather than explicit rules. This paper introduces CRISP, a framework that reconstructs the last-layer activation vector (LLAV) of binary neural teachers as Tsetlin Machine (TM) clauses. CRISP sign-binarizes the teacher's penultimate pre-logit activations, and assigns one Individual TM (ITM) to each LLAV neuron. Each reconstructed hidden bit is represented by propositional clauses over Booleanized input features, which gives a direct symbolic trace from named input thresholds to a named teacher neuron. CRISP is evaluated on MNIST, KMNIST, FashionMNIST (FMNIST), SVHN, and CIFAR10 using a BinaryConnect convolutional neural network (BCCNN) teacher and a fully binary neural network (BNN) teacher, with an additional study on binary thresholding, thermometer encoding, and quartile binning at multiple bit depths. The results show that LLAV sign-binarization does not reduce teacher-head accuracy in the tested BNN setting, while ITM reconstruction error is the main limiting factor. Quartile one-bit Booleanization gives the strongest reconstruction fidelity on SVHN at 87.52% test fidelity and is competitive on CIFAR10, and the reconstructed LLAV preserves 78.41% teacher-head accuracy on FMNIST. Pooled clause-evidence visualizations show that the learned ITM literals concentrate on the object region in centered benchmarks. CRISP therefore provides a clause-level route for inspecting the final hidden representation of binary neural teachers.
CommentsPreprint for an accepted paper at International Symposium of the Tsetlin Machine (ISTM) 2026 Conference
