IEEE Northern Jersey Section, SMC Chapter Seminar on Domain Adaptation via Enhanced Subspace Distribution Matching and Generative Adversarial Distribution Matching

---

Domain Adaptation via Enhanced Subspace Distribution Matching and Generative Adversarial Distribution Matching

 

Siya Yao, Ph.D. Candidate

Department of Control Science and Engineering

Tongji University, Shanghai, China

Time: 10am, Tuesday, October 5, 2021

Place: ECEC 207, New Jersey Institute of Technology

Virtual: https://njit.webex.com/meet/zhou

 

Abstract: In the field of computer vision, without sufficient labeled images, it is challenging to train an accurate model. Domain adaptation aims to accomplish tasks on unlabeled target data by learning and transferring knowledge from related source domains. In order to learn a discriminative and domain-invariant model, a critical step is to align source and target data well and thus reduce their distribution divergence. Various traditional and deep methods are developed to deal with such issue. Most prior traditional approaches merely reduce subspace conditional or marginal distribution differences between domains but entirely ignoring label dependence information of source data in subspace. We propose a novel approach of domain adaptation, called enhanced subspace distribution matching (ESDM), which makes good use of label information to enhance the distribution matching between the source and target domains in a shared subspace. As for deep methods, many methods apply adversarial learning to diminish cross-domain distribution difference. Generative adversarial network loss is widely used in adversarial adaptation learning methods. However, it becomes difficult to decline distribution difference if generator or discriminator in GAN fails to work as expected and degrades its performance. To solve such cross-domain classification problems, we put forward a novel adaptation framework called generative adversarial distribution matching (GADM). In GADM, we improve the objective function by taking cross-domain discrepancy distance into consideration and further minimize the difference through the competition between a generator and discriminator, thereby greatly decreasing cross-domain distribution difference. Experimental results and comparison with several state-of-the-art methods verify both ESDM and GADM’s superiority in across-domain image classification.

 

Siya Yao received her B.S. degree in Automation, from Donghua University, Shanghai, China in 2017. She is currently pursuing a Ph.D. degree in Control Science and Engineering with the Department of Control Science and Engineering, Tongji University, Shanghai, China. Since 2019, she has been working as a joint Ph.D. Student with the Department of ECE, New Jersey Institute of Technology, Newark, NJ, USA. Her research interests are in transfer learning and anomaly detection.

 

Contact: Prof. Mengchu Zhou, zhou@njit.edu

---

---

---

---

10-11:10 Seminar and Social networking

---