Timezone: »

Implicit Class-Conditioned Domain Alignment for Unsupervised Domain Adaptation
Xiang Jiang · Qicheng Lao · Stan Matwin · Mohammad Havaei

Thu Jul 16 07:00 AM -- 07:45 AM & Thu Jul 16 06:00 PM -- 06:45 PM (PDT) @ Virtual #None

We present an approach for unsupervised domain adaptation—with a strong focus on practical considerations of within-domain class imbalance and between-domain class distribution shift—from a class-conditioned domain alignment perspective. Current methods for class-conditioned domain alignment aim to explicitly minimize a loss function based on pseudo-label estimations of the target domain. However, these methods suffer from pseudo-label bias in the form of error accumulation. We propose a method that removes the need for explicit optimization of model parameters from pseudo-labels. Instead, we present a sampling-based implicit alignment approach, where the sample selection is implicitly guided by the pseudo-labels. Theoretical analysis reveals the existence of a domain-discriminator shortcut in misaligned classes, which is addressed by the proposed approach to facilitate domain-adversarial learning. Empirical results and ablation studies confirm the effectiveness of the proposed approach, especially in the presence of within-domain class imbalance and between-domain class distribution shift.

Author Information

Xiang Jiang (Imagia, Dalhousie University)
Qicheng Lao (Mila; Imagia)
Stan Matwin (Dalhousie University)
Mohammad Havaei (Imagia)