CAST: Closed-form Analytic Semantic Transfer for Zero-Shot Classifier Extension

Researchers introduced CAST (Closed-form Analytic Semantic Transfer), a training-free, image-free framework that extends pre-trained classifiers to unseen classes via weight injection, achieving performance comparable to few-shot methods on standard zero-shot learning benchmarks without requiring target-distribution examples. The framework includes a computable semantic extrapolation residual that guides dataset curation and benchmark design.

arXiv:2608.13751v1 Announce Type: new Abstract: Large pre-trained models have become foundational components of modern machine learning systems. Yet adapting these models to novel categories typically requires examples from the target distribution. In many domains, however, such data are unavailable. Zero-shot learning ZSL permits recognition under these limitations through relying on auxiliary semantic information such as textual descriptions. We introduce CAST Closed-form Analytic Semantic Transfer , a training-free, image-free framework for extending a pre-trained classifier to previously unseen classes through weight injection. We provide a theoretical foundation for CAST and derive a finite-sample error decomposition that identifies the \emph{semantic extrapolation residual} $\rho u$. The residual is a computable, model-agnostic measure and provides a principled criterion for dataset curation and benchmark design. Experiments on standard zero-shot learning benchmarks demonstrate that CAST matches or exceeds existing image-free approaches and approaches the performance of few-shot adaptation methods, while requiring neither iterative optimization nor examples from the target distribution.