Upper Bound of Bayesian Generalization Error in Partial Concept Bottleneck Model (CBM): Partial CBM outperforms naive CBM
This provides a theoretical foundation for PCBM, addressing a gap in understanding generalization in singular statistical models, though it is incremental as it focuses on a specific linear case.
The paper tackles the problem of theoretical generalization error in Partial Concept Bottleneck Models (PCBM), which use partially observed concepts to improve interpretability without sacrificing performance, and finds that PCBM reduces Bayesian generalization error compared to full-observed CBM in a linear three-layer architecture.
Concept Bottleneck Model (CBM) is a methods for explaining neural networks. In CBM, concepts which correspond to reasons of outputs are inserted in the last intermediate layer as observed values. It is expected that we can interpret the relationship between the output and concept similar to linear regression. However, this interpretation requires observing all concepts and decreases the generalization performance of neural networks. Partial CBM (PCBM), which uses partially observed concepts, has been devised to resolve these difficulties. Although some numerical experiments suggest that the generalization performance of PCBMs is almost as high as that of the original neural networks, the theoretical behavior of its generalization error has not been yet clarified since PCBM is singular statistical model. In this paper, we reveal the Bayesian generalization error in PCBM with a three-layered and linear architecture. The result indcates that the structure of partially observed concepts decreases the Bayesian generalization error compared with that of CBM (full-observed concepts).