TACIT: A Target-Agnostic Feature Disentanglement Framework for Cross-Domain Text Classification
This addresses a practical limitation in cross-domain text classification for applications where target domains are unknown, though it is incremental as it builds on existing feature disentanglement methods.
The paper tackles the problem of cross-domain text classification when target domain data is unavailable, proposing TACIT to disentangle robust and unrobust features using Variational Auto-Encoders and feature distillation, achieving results comparable to state-of-the-art baselines using only source domain data.
Cross-domain text classification aims to transfer models from label-rich source domains to label-poor target domains, giving it a wide range of practical applications. Many approaches promote cross-domain generalization by capturing domain-invariant features. However, these methods rely on unlabeled samples provided by the target domains, which renders the model ineffective when the target domain is agnostic. Furthermore, the models are easily disturbed by shortcut learning in the source domain, which also hinders the improvement of domain generalization ability. To solve the aforementioned issues, this paper proposes TACIT, a target domain agnostic feature disentanglement framework which adaptively decouples robust and unrobust features by Variational Auto-Encoders. Additionally, to encourage the separation of unrobust features from robust features, we design a feature distillation task that compels unrobust features to approximate the output of the teacher. The teacher model is trained with a few easy samples that are easy to carry potential unknown shortcuts. Experimental results verify that our framework achieves comparable results to state-of-the-art baselines while utilizing only source domain data.