Instruction and Demonstration-based service attribute Mining mechanism for textual data

Attribute-based access control is fundamentally dependent on the service attribute calibration of object sources. Traditionally, the calibration of service attribute for textual data has been primarily reliant on human experts and machine learning methods, yet the efficiency and few-shot ability are insufficient. Thus, a service attribute mining mechanism for textual data, termed IDM, has been introduced. This mechanism addresses the aforementioned challenges by transforming the extraction of service attribute, previously a sequence-calibrated problem, into a controllable-generation problem through instruction learning and in-context learning. The service attribute extraction component of the proposed mechanism achieved an optimal average F1 score in few-shot experiments on the CoNLL2003 dataset, surpassing the baseline model. This positions the mechanism as state-of-the-art.