Data Integrity and Artificial Reasoning.

Data integrity viewed from different perspectives has overlapping and distinct definitions. In general, data integrity is the maintaining of accuracy and consistency over time as well as the assurance of the accuracy and consistency within the data. There is also the overlap of data integrity and data quality. For data to have integrity it should also ideally be whole, undivided, sound, consistent and without corruption. These ideas are the essence of data integrity. For the perspective of the research within Artificial Reasoning, data integrity includes having complete, valid, timely and even unique data. Also, the concept of not just limiting the uncertainty of data but capturing the type of uncertainty is important, particularly for tasks such as decision making. A key area of research within Artificial Reasoning is causal reasoning. Inferring using causality has many strengths. One is the ability to connect useful information within and across modalities. With this comes the ability to reduce the redundancy of data to improve results. This vision paper will present the extension of the Artificial Reasoning research in causality that acknowledges the impact of data integrity and potential solutions. In this paper, we discuss the ideas that bridge data integrity and artificial reasoning, specifically causal reasoning. Specifically, a new causal framework with three main components is proposed. These components would leverage learned representations, handle uncertainty, and integrate data to maintain relevancy and limit unnecessary redundancy. In addition, we present the causal framework for this concept.