Memory models for skill and experience

• Main Author: Atif Saleem
• Other Authors: Tan Ah Hwee
• Format: Final Year Project
• Language: English
• Published: 2014
• Subjects: DRNTU::Engineering::Computer science and engineering::Computing methodologies::Pattern recognition
• Online Access: http://hdl.handle.net/10356/59988
• Institution: Nanyang Technological University

Episodic memory is the collection of past personal experiences that occurred at a particular time and place. Episodic memory is closely tied to emotion, and it is intrinsic human behaviour to indulge in episodes of the past when being in a particular affective state. It is most common among the aged community, who often draw upon their memories for the purposes of comfort and relaxation. This project aims to digitally simulate episodic memory storage, retrieval and playback for the elderly to amplify their fading memories. Traditional database methods are unable to understand the complex relations between events and episodes and are hence unsuitable to meet the objectives of this project. Rather, this project employs EM-ART, a self- organizing neural network that closely mimics the attributes and behaviour of human episodic memory. In this way, the model is able to perform complex sequential learning tasks. The model was trained with test cases shown in appendix B for experiments, and the results from a series of test cases were evaluated. The results show that the model is able to learn complex relations between events and retrieve episodes as a chunk with imperfect or partial cues effectively and appropriately.