Virtual Adversarial Training in Feature Space to Improve Unsupervised Video Domain Adaptation
This work addresses domain adaptation challenges for video analysis, but it is incremental as it builds on existing methods like TA³N.
The paper tackled the problem of improving unsupervised video domain adaptation by applying virtual adversarial training directly to feature vectors instead of pixel space, and addressing instability in entropy minimization and decision-boundary iterative refinement training. The result was that adding these techniques to the TA³N model maintained competitive results or outperformed prior art in multiple tasks.
Virtual Adversarial Training has recently seen a lot of success in semi-supervised learning, as well as unsupervised Domain Adaptation. However, so far it has been used on input samples in the pixel space, whereas we propose to apply it directly to feature vectors. We also discuss the unstable behaviour of entropy minimization and Decision-Boundary Iterative Refinement Training With a Teacher in Domain Adaptation, and suggest substitutes that achieve similar behaviour. By adding the aforementioned techniques to the state of the art model TA$^3$N, we either maintain competitive results or outperform prior art in multiple unsupervised video Domain Adaptation tasks