Abstract Attention-based neural models were employed to detect the different aspects and sentiment polarities of the same target in targeted aspectbased sentiment analysis (TABSA). However, existing methods do not specififically pre-train reasonable embeddings for targets and aspects in TABSA. This may result in targets or aspects having the same vector representations in different contexts and losing the contextdependent information. To address this problem, we propose a novel method to refifine the embeddings of targets and aspects. Such pivotal embedding refifinement utilizes a sparse coeffificient vector to adjust the embeddings of target and aspect from the context. Hence the embeddings of targets and aspects can be refifined from the highly correlative words instead of using context-independent or randomly initialized vectors. Experiment results on two benchmark datasets show that our approach yields the state-of-the-art performance in TABSA task