Metastimuli: An Introduction to PIMS Filtering

Abstract

A system design for correlating information stimuli and a user’s personal information management system (PIMS) is introduced. This is achieved via a deep learning classifier for textual data, a recently developed PIMS graph information architecture, and a principle component analysis (PCA) reduction thereof. The system is designed to return unique and meaningful signals from incoming textual data in or near realtime. The classifier uses a recurrent neural network to determine the location of a given atom of information in the user’s PIMS. PCA reduction of the PIMS graph to \(\mathbb {R}^m\), with m the actuator (haptic) dimensionality, is termed a PIMS filter. Demonstrations are given of the classifier and PIMS filter. The haptic stimuli, then, are correlated with the user’s PIMS and are therefore termed “metastimuli.” Applications of this system include educational environments, where human learning may be enhanced. We hypothesize a metastimulus bond effect on learning that has some support from the analogous haptic bond effect. A study is outlined to test this hypothesis.