Target Domain Data induces Negative Transfer in Mixed Domain Training with Disjoint Classes
This addresses a practical issue for machine learning practitioners using domain adaptation, revealing a counterintuitive negative effect that is incremental but important for real-world applications.
The paper tackles the problem of negative transfer in mixed domain training when target domain data is included but has disjoint classes from surrogate domains, showing that this inclusion significantly decreases performance, sometimes below random, across over 25 domain shifts.
In practical scenarios, it is often the case that the available training data within the target domain only exist for a limited number of classes, with the remaining classes only available within surrogate domains. We show that including the target domain in training when there exist disjoint classes between the target and surrogate domains creates significant negative transfer, and causes performance to significantly decrease compared to training without the target domain at all. We hypothesize that this negative transfer is due to an intermediate shortcut that only occurs when multiple source domains are present, and provide experimental evidence that this may be the case. We show that this phenomena occurs on over 25 distinct domain shifts, both synthetic and real, and in many cases deteriorates the performance to well worse than random, even when using state-of-the-art domain adaptation methods.