Variational interaction information maximization for cross-domain disentanglement도메인 간 은닉 공간 분리를 위한 변형 상호 작용 정보 최대화

Cited 0 time in webofscience Cited 0 time in scopus
  • Hit : 168
  • Download : 0
Cross-domain disentanglement is the problem of learning representations partitioned into domain-invariant and domain-specific representations, which is a key to successful domain transfer or measuring semantic distance between two domains. Grounded in information theory, we cast the simultaneous learning of domain-invariant and domain-specific representations as a joint objective of multiple information constraints, which does not require adversarial training or gradient reversal layers. We derive a tractable bound of the objective and propose a generative model named Interaction Information Auto-Encoder. Our approach reveals insights on the desirable representation for cross-domain disentanglement and its connection to Variational Auto-Encoder. We demonstrate the validity of our model in the image-to-image translation and the cross-domain retrieval tasks. We further show that our model achieves the state-of-the-art performance in the zero-shot sketch based image retrieval task, even without external knowledge.
Advisors
Kim, Kee-Eungresearcher김기응researcher
Description
한국과학기술원 :전산학부,
Publisher
한국과학기술원
Issue Date
2020
Identifier
325007
Language
eng
Description

학위논문(석사) - 한국과학기술원 : 전산학부, 2020.8,[iv, 18 p :]

Keywords

Cross-domain disentanglement▼aInformation theory▼aInteraction information▼aInvariant representation▼aLatent variable▼aGenerative model▼aVariational inference; 도메인 간 은닉공간 분리▼a정보이론▼a상호 작용 정보▼a불변 표현▼a은닉변수▼a생성모델▼a변형 추론

URI
http://hdl.handle.net/10203/285213
Link
http://library.kaist.ac.kr/search/detail/view.do?bibCtrlNo=947947&flag=dissertation
Appears in Collection
CS-Theses_Master(석사논문)
Files in This Item
There are no files associated with this item.

qr_code

  • mendeley

    citeulike


rss_1.0 rss_2.0 atom_1.0