A Classifier That Teaches Itself: Self-Improving, Frozen-gate Training (SIFT) for Dynamic Document Classification

自我进化的分类器：用于动态文档分类的“自我提升、冻结门控训练”（SIFT）

Abstract: Document classification is a solved problem in the laboratory and an unsolved one in the enterprise. The blocker is rarely model architecture; it is the labeling project that must precede a model and the institutional fear of letting a model retrain itself once one exists. We present SIFT (Self-Improving, Frozen-gate Training), a dynamic classifier service, which attacks both.

摘要： 文档分类在实验室环境下已是一个被解决的问题，但在企业应用中却依然困难重重。阻碍因素往往并非模型架构本身，而是模型部署前必须进行的标注工程，以及机构对于允许模型自动重训练的顾虑。我们提出了 SIFT（自我提升、冻结门控训练），这是一种动态分类器服务，旨在同时解决上述两个痛点。

SIFT serves classification from a deliberately cheap, CPU-bound pipeline, a SPLADE sparse encoder feeding a LightGBM head, and escalates only the low-confidence minority of pages to an LLM judge. The judge’s verdicts are written back into a labeled corpus, so the expensive model continuously teaches the cheap one: the escalation rate falls, the corpus grows from production traffic rather than from an up-front annotation effort, and accuracy compounds with use.

SIFT 通过一种刻意设计的低成本、CPU 密集型流水线提供分类服务，即利用 SPLADE 稀疏编码器配合 LightGBM 分类头，仅将置信度较低的少数页面升级给大语言模型（LLM）裁判进行判断。裁判的判定结果会被回写到标注语料库中，从而实现昂贵的模型持续“教导”廉价模型：随着使用，升级率不断下降，语料库通过生产流量而非预先标注工作不断增长，分类准确率也随之累积提升。

Onboarding a new document family requires only a declarative bundle, label space, anchor phrases, and a judge glossary, not a labeling project. The harder problem is safety: an autonomously retraining classifier can silently regress. SIFT resolves this with a two-part promote gate, a critical-label F1 regression check plus a frozen golden regression set the model is never trained on, either of which vetoes promotion. This turns “retrain monthly without a human” from reckless into routine.

引入新的文档类别仅需一个声明式配置包、标签空间、锚点短语和裁判词汇表，无需进行繁琐的标注工程。更棘手的问题在于安全性：自主重训练的分类器可能会发生隐性性能退化。SIFT 通过两阶段的“提升门控”（promote gate）机制解决了这一问题：即关键标签的 F1 回归检查，以及一个模型从未参与训练的“冻结黄金回归集”。这两者中的任何一个触发警报，都会否决模型的升级。这使得“每月无需人工干预的自动重训练”从一种鲁莽的行为变成了常规操作。

We describe the architecture, the self-feeding corpus loop, the frozen-gate promotion mechanism, and an illustrative multi-domain deployment, and we discuss the economics of a classifier whose marginal labeling cost trends toward zero.

我们详细描述了该架构、自供给语料库循环、冻结门控提升机制，以及一个多领域部署的示例，并探讨了这种边际标注成本趋近于零的分类器的经济效益。