Heterogeneous Transfer Learning via Deep Matrix Completion with Adversarial Kernel Embedding

• DOI: 10.1609/aaai.v33i01.33018602
• Publication Date: 2019-08-20T07:47:57Z
• Authors: Haoliang Li; Sinno Jialin Pan; Renjie Wan; Alex C. Kot
• Keywords: Transfer of learning; Kernel (algebra); Matrix Completion; Feature (linguistics); Matrix (chemical analysis)

Heterogeneous Transfer Learning (HTL) aims to solve transfer learning problems where a source domain and target are of heterogeneous types features. Most existing HTL approaches either explicitly learn feature mappings between the domains or implicitly reconstruct cross-domain features based on matrix completion techniques. In this paper, we propose new method deep framework, kernel embedding distributions is trained in an adversarial manner for across domains. We conduct extensive experiments two different vision tasks demonstrate effectiveness our proposed compared with number baseline methods.