Domain Adaptation On Manifolds

Domain adaptation aims to remedy the loss in classification performance that often occurs due to domain shifts between training and testing datasets.

Domain adaptation on manifolds. In a recent work we proposed to view the data through the lens of covariance matrices and presented a method for domain adaptation using parallel transport on the cone manifold of symmetric positive definite matrices. As a pre processing step our approach can also be combined with existing domain adaptation approaches to learn a common feature space for all input domains. Meanwhile only a small fraction of the target instances. The target data distribution is under certain unknown transformation of the source data distribution.

Specifically we propose to use low dimensional manifold to. Domain adaptation methods on grassmann manifolds are among the most popular including geodesic subspace sampling and geodesic flow kernel. 05 06 2020 by pengfei wei et al. Cheng l pan sj.

Domain adaptation techniques which focus on adapting models between distributionally different domains are rarely explored in the video recognition area due to the significant spatial and. Semi supervised domain adaptation on manifolds. Reducing domain divergence is a key step in transfer learning problems. Request pdf semi supervised domain adaptation on manifolds in real life problems the following semi supervised domain adaptation scenario is often encountered.

Reducing domain divergence is a key step in transfer learning problems. 0 share. Domain adaptation on the statistical manifold mahsa baktashmotlagh1 3 mehrtash t. Existing works focus on the minimization of global domain divergence.

Existing works focus on the minimization of global domain divergence. We currently put a special focus on the problem of domain adaptation. In this paper we take the local divergence of subdomains into account in transfer. This problem is known as the dataset bias attributed to variations across datasets.

In real life problems the following semi supervised domain adaptation scenario is often encountered. Subdomain adaptation with manifolds discrepancy alignment. We have full access to some source data which is usually very large. However two domains may consist of several shared subdomains and differ from each other in each subdomain.