Provable Adaptation across Multiway Domains via Representation Learning
This work addresses the problem of adapting to unseen domains in multiway settings for machine learning practitioners, offering theoretical guarantees and empirical validation, though it is incremental in extending existing domain adaptation frameworks.
The paper tackles zero-shot domain adaptation across multi-dimensional domains with limited training data, proposing a model with a domain-invariant representation and domain-specific low-rank tensor layers, and provides the first finite-sample guarantees with explicit error bounds, supported by experiments on MNIST and fiber sensing datasets.
This paper studies zero-shot domain adaptation where each domain is indexed on a multi-dimensional array, and we only have data from a small subset of domains. Our goal is to produce predictors that perform well on \emph{unseen} domains. We propose a model which consists of a domain-invariant latent representation layer and a domain-specific linear prediction layer with a low-rank tensor structure. Theoretically, we present explicit sample complexity bounds to characterize the prediction error on unseen domains in terms of the number of domains with training data and the number of data per domain. To our knowledge, this is the first finite-sample guarantee for zero-shot domain adaptation. In addition, we provide experiments on two-way MNIST and four-way fiber sensing datasets to demonstrate the effectiveness of our proposed model.