Cross-domain Adaptation with Discrepancy Minimization for Text-independent Forensic Speaker Verification. (arXiv:2009.02444v1 [eess.AS]) http://arxiv.org/abs/2009.02444