arXiv · cs.AI · Mochamad Rizal Fauzan, Naila Nursyifa, Irgi Surya2026-11-03 13:00 · 8 小时前模型评分36

Halal-NeSyNet：用子句级神经符号交叉注意力审计食品配料并做风险分级

Halal-NeSyNet: Clause-Scoped Neuro-Symbolic Cross-Attention for Automated Food Ingredient Auditing and Risk-Aware Halal Logistics Triage

导读

Halal-NeSyNet 将 DistilBERT 编码器与 16 维规则符号向量通过交叉注意力与门控融合结合，用于识别同一子句内植物性限定词误清另一子句未限定添加剂的问题。

分类置信度

中 · 待人工复核

来源级别：媒体与个人

应用领域：制造与工业、交通与物流

事件类型：研究发布

证据摘录

“We propose Halal-NeSyNet, which couples a DistilBERT encoder with a 16-dimensional rule-derived symbolic vector through cross-attention and gated fusion, trained with an asymmetric false-Halal risk loss.”
“Because labels and features come from one rule engine and 91.3 % of the items are templated, the results measure fidelity to a rule policy, not religious certification.”

来源：arXiv · cs.AI · arxiv.org