Search papers, labs, and topics across Lattice.
This paper introduces the Chain of Operators (CHOP) framework, which leverages a frozen In-Context Operator Network (ICON) to effectively tackle out-of-distribution (OOD) operator tasks without requiring parameter updates. By constructing a chain of interpretable elementary transformations alongside the ICON, CHOP demonstrates a significant reduction in relative inference error compared to direct ICON evaluations on scalar conservation laws and mean-field control problems. Notably, the framework shows that a chain built for one family of partial differential equations (PDEs) can generalize to another, highlighting the potential for shared mechanisms across different operator tasks.
CHOP achieves superior performance on OOD tasks by utilizing a frozen ICON, revealing that interpretability and generalization can coexist in operator learning.
Neural operators approximate mappings between function spaces, but often generalize poorly to other operators and usually require fine-tuning or retraining. In-Context Operator Networks (ICON) addresses this issue by prompting the model with numerical context so that the model learns specific operators from prompts and adapt to different operators without fine-tuning. However, ICON may still fail to generalize to out-of-distribution (OOD) operator tasks. Inpired by the success of harness engineering of Large Language models (LLMs), we introduce Chain of Operators (CHOP), a framework that harness a frozen ICON to OOD operator tasks without updating its parameters. Specifically, CHOP constructs a chain of operators consisting of explicit elementary transformations and the frozen ICON. Experiments on a scalar conservation law and a mean-field control problem show that CHOP reduces relative inference error over direct ICON evaluation, while each operator in the chain remains interpretable and in closed form. A chain constructed on one PDE family further generalizes to a different family, indicating shared mechanisms across harness systems.
NUS