Recent reports suggest that a generic supervised deep CNN model trained on a large-scale dataset reduces, but does not remove, dataset bias on a standard benchmark. Fine-tuning deep models in a new domain can require a significant amount of data, whi
ch for many applications is simply not available. We propose a new CNN architecture which introduces an adaptation layer and an additional domain confusion loss, to learn a representation that is both semantically meaningful and domain invariant. We additionally show that a domain confusion metric can be used for model selection to determine the dimension of an adaptation layer and the best position for the layer in the CNN architecture. Our proposed adaptation method offers empirical performance which exceeds previously published results on a standard benchmark visual domain adaptation task.
In the late stage of the evolution of a pion system in high-energy heavy-ion collisions when pions undergo multiple scatterings, the quantum transport of the interfering pair of identical pions plays an important role in determining the characteristi
cs of the Hanbury-Brown-Twiss (HBT) interference. We study the quantum transport of the interfering pair using the path-integral method, in which the evolution of the bulk matter is described by relativistic hydrodynamics while the paths of the two interfering pions by test particles following the fluid positions and velocity fields. We investigate in addition the effects of secondary pion sources from particle decays, for nuclear collisions at AGS and RHIC energies. We find that quantum transport of the interfering pair leads to HBT radii close to those for the chemical freeze-out configuration. Particle decays however lead to HBT radii greater than those for the chemical freeze-out configuration. As a consequence, the combined effects give rise to HBT radii between those extracted from the chemical freeze-out configuration and the thermal freeze-out configuration. Proper quantum treatments of the interfering pairs in HBT calculations at the pion multiple scattering stage are important for our understanding of the characteristics of HBT interferometry in heavy-ion collisions.
