This document describes the submissions of STAR-LAB (the Speech Technology and Research Laboratory at SRI Interna- tional) to the open-set condition of the IberSPEECH-RTVE 2018 Speaker Diarization Challenge. The core components of the submissions included noise-robust speech activity detec- tion, speaker embeddings for initializing diarization with do- main adaptation, and variational Bayes (VB) diarization using a DNN bottleneck i-vector subspaces.