Multi-Adversarial Domain Adaptation

TLDR

Deep domain adaptation uses adversarial learning to reduce distribution gaps, but existing single‑discriminator methods fail to capture complex multimode structures. The paper proposes MADA, a multi‑adversarial domain adaptation method that captures multimode structures for fine‑grained alignment using multiple discriminators. MADA is trained via stochastic gradient descent, computing gradients by back‑propagation in linear time. Experiments show that MADA outperforms state‑of‑the‑art methods on standard domain adaptation benchmarks.

Abstract

Recent advances in deep domain adaptation reveal that adversarial learning can be embedded into deep networks to learn transferable features that reduce distribution discrepancy between the source and target domains. Existing domain adversarial adaptation methods based on single domain discriminator only align the source and target data distributions without exploiting the complex multimode structures. In this paper, we present a multi-adversarial domain adaptation (MADA) approach, which captures multimode structures to enable fine-grained alignment of different data distributions based on multiple domain discriminators. The adaptation can be achieved by stochastic gradient descent with the gradients computed by back-propagation in linear-time. Empirical evidence demonstrates that the proposed model outperforms state of the art methods on standard domain adaptation datasets.