Compile by Training: Turning Natural-Language Specifications into Local Neural Functions

• 类型：arxiv
• 标识：2609.04199
• 链接：https://arxiv.org/abs/2609.04199
• 主分类：engineering
• 形态：method
• 被引：0
• 被引来源：Semantic Scholar + OpenAlex
• S2被引：0
• OpenAlex被引：0
• 影响力被引：0
• TLDR：Many recurring text functions are easy to describe but difficult to implement with rules, while calling a large remote model for every input introduces repeated cost, latency, and dependency on a provider. We present compile by training, which turns a natural-language specification into a reusable neural function. At compile time, teacher models generate task-specific examples that are used to train a small adapter for a compact interpreter. The resulting function runs without the teachers and can be stored, versioned, and composed like ordinary software. On FuzzyBench-Hard, a subset on which
• OpenAlex ID：W7208734651
• OpenAlex DOI：10.48550/arxiv.2609.04199
• DOI：10.48550/arxiv.2609.04199
• DOI来源：OpenAlex
• 开放获取：green
• 开放获取链接：https://doi.org/10.48550/arxiv.2609.04199
• OpenAlex更新：2026-09-06
• 待LLM分类：否
• 标题中文：Compile by Training：将自然语言规约转化为本地神经函数
• TLDR中文：许多重复出现的文本函数易于描述却难以用规则实现；而为每个输入调用大型远程模型会带来重复开销、延迟与对服务方的依赖。我们提出 compile by training，将自然语言规约转化为可复用的神经函数。在编译时，教师模型生成任务专属样本，用于为精简解释器训练一个小型适配器。生成的函数可在没有教师模型的情况下运行，并能像普通软件一样被存储、版本化管理与组合。在 FuzzyBench-Hard 这一子集上……
• 来源文件：
• /inbox/tom/_candidates/2026-09-04-agent-rag-longcontext-candidates.json
• [OpenAlex backfill]
• [S2 enrich]