Probing Neural Combinatorial Optimization Models
This work addresses the need for deeper insights into NCO models for researchers and stakeholders, representing a first systematic attempt to interpret these black-box models, though it is incremental as it builds on existing probing techniques.
The paper tackles the problem of interpreting black-box neural combinatorial optimization (NCO) models by investigating their representations through probing tasks, introducing a novel tool called Coefficient Significance Probing (CS-Probing) that reveals insights such as inductive biases and key embedding dimensions, and demonstrates practical improvements like enhancing model generalization with minor code modifications.
Neural combinatorial optimization (NCO) has achieved remarkable performance, yet its learned model representations and decision rationale remain a black box. This impedes both academic research and practical deployment, since researchers and stakeholders require deeper insights into NCO models. In this paper, we take the first critical step towards interpreting NCO models by investigating their representations through various probing tasks. Moreover, we introduce a novel probing tool named Coefficient Significance Probing (CS-Probing) to enable deeper analysis of NCO representations by examining the coefficients and statistical significance during probing. Extensive experiments and analysis reveal that NCO models encode low-level information essential for solution construction, while capturing high-level knowledge to facilitate better decisions. Using CS-Probing, we find that prevalent NCO models impose varying inductive biases on their learned representations, uncover direct evidence related to model generalization, and identify key embedding dimensions associated with specific knowledge. These insights can be potentially translated into practice, for example, with minor code modifications, we improve the generalization of the analyzed model. Our work represents a first systematic attempt to interpret black-box NCO models, showcasing probing as a promising tool for analyzing their internal mechanisms and revealing insights for the NCO community. The source code is publicly available.