Information Retrieval (IR) is the task of obtaining pieces of data (such as\ndocuments) that are relevant to a particular query or need from a large\nrepository of information. IR is a valuable component of several downstream\nNatural Language Processing (NLP) tasks. Practically, IR is at the heart of\nmany widely-used technologies like search engines. While probabilistic ranking\nfunctions like the Okapi BM25 function have been utilized in IR systems since\nthe 1970's, modern neural approaches pose certain advantages compared to their\nclassical counterparts. In particular, the release of BERT (Bidirectional\nEncoder Representations from Transformers) has had a significant impact in the\nNLP community by demonstrating how the use of a Masked Language Model trained\non a large corpus of data can improve a variety of downstream NLP tasks,\nincluding sentence classification and passage re-ranking. IR Systems are also\nimportant in the biomedical and clinical domains. Given the increasing amount\nof scientific literature across biomedical domain, the ability find answers to\nspecific clinical queries from a repository of millions of articles is a matter\nof practical value to medical professionals. Moreover, there are\ndomain-specific challenges present, including handling clinical jargon and\nevaluating the similarity or relatedness of various medical symptoms when\ndetermining the relevance between a query and a sentence. This work presents\ncontributions to several aspects of the Biomedical Semantic Information\nRetrieval domain. First, it introduces Multi-Perspective Sentence Relevance, a\nnovel methodology of utilizing BERT-based models for contextual IR. The system\nis evaluated using the BioASQ Biomedical IR Challenge. Finally, practical\ncontributions in the form of a live IR system for medics and a proposed\nchallenge on the Living Systematic Review clinical task are provided.\n