将学习问题编译为语言模型的适应程序

中文摘要 AI 辅助

模型适应通常由固定配方控制，尽管不同的更新程序可能产生截然不同的行为结果。我们引入了适应编译，它将模型应在何处、如何以及何种程度上进行适应重新定义为一个联合预测与决策问题。编译器不是为每个学习片段重新搜索候选程序，而是从先前的适应中学习，以预测候选程序上的向量值反事实响应面——它们在获取、迁移、有界性和保持方面的预期效果——并在适应开始前选择一个程序。由于这种预测几何结构捕捉了多种行为后果，而非单一赢家或标量分数，因此可以在不同下游优先级下重用，而无需重新训练。在五种学习类型中，首选程序在不同片段间存在显著差异，且这种差异可从适应前信息中预测。在Llama-3.1-8B上，编译器选择的程序接近穷举搜索，同时优于全局和特定目标的默认设置。在Gemma-2-9B上的复现保留了程序异质性和选择余量，但表明利用这一余量需要在偏离强默认设置时考虑不确定性。综合来看，这些结果表明，适应搜索可以在相关学习问题间摊销，将先前的适应经验转化为决定未来学习应如何发生的基础。

英文摘要

Model adaptation is typically governed by a fixed recipe, even though different update programs can produce substantially different behavioral outcomes. We introduce adaptation compilation, which reframes where, how, and to what extent a model should adapt as a joint prediction and decision problem. Rather than searching over candidate programs anew for each learning episode, a compiler learns from prior adaptations to predict a vector-valued counterfactual response surface over candidate programs---their expected effects on acquisition, transfer, boundedness, and preservation---and selects a program before adaptation begins. Because this predicted geometry captures multiple behavioral consequences rather than a single winner or scalar score, it can be reused under different downstream priorities without retraining. Across five learning types, preferred programs vary meaningfully across episodes, and this variation is predictable from pre-adaptation information. On Llama-3.1-8B, compiler-selected programs approach exhaustive search while outperforming global and objective-specific defaults. Replication on Gemma-2-9B preserves program heterogeneity and selection headroom, but shows that exploiting this headroom requires accounting for uncertainty when departing from strong defaults. Together, these results show that adaptation search can be amortized across related learning problems, turning prior adaptation experience into a basis for deciding how future learning should occur.

发表机构

• Yale University（耶鲁大学）

机构由 AI 辅助整理，请以论文原文为准。