News & Events

Share        

ICAA 2017 Best Paper Award

Published on: 22-Aug-2017

A research through the collaboration between NTU and Western Sydney University (WSU) received the Best Paper award at the 5th International Conference on Ageless Ageing (ICAA 2017), which was held in the beautiful campus of Tsinghua University, Beijing, China in July. This research explores how to model human autobiographical memory using a neural network and further attempts to quantitatively study memory loss in each phase of memory encoding, storage and retrieval.


The members of the research team are (from left) Prof Miao Chunyan, Prof Tan Ah-Hwee, Dr Wang Di and the overseas collaborator, Dr Ahmed Moustafa (not in photo).

The research team consists of Dr WANG Di, a research fellow of the Joint NTU-UBC Research Centre of Excellence in Active Living for the Elderly (LILY), Dr TAN Ah Hwee, a professor of School of Computer Science and Engineering (SCSE) and PI of LILY, Dr MIAO Chunyan, a professor of SCSE and director of LILY, and the overseas collaborator, Dr Ahmed MOUSTAFA, a senior lecturer of School of Social Sciences and Psychology, WSU.

The team’s awarded work presents a neural network, named Autobiographical Memory-Adaptive Resonance Theory network (AM-ART), to model autobiographical memory, which is a core component of human brain and plays an important role in self-identification. Cognitive neuroscientists have constructed conceptual models to explain the functionalities of autobiographical memory and describe its dynamics. Some of these models have already been physiologically supported by neural imaging. However, most existing computational autobiographical memory models do not distinguish from other long term memory models. Specifically, during model design, they do not take the specific features of autobiographical memory and its unique encoding, storage and retrieval processes into account. The awarded research work introduces AM-ART, which is consistent with a widely adopted cognitive model that has been supported by neural imaging, in terms of both the neural network structure and the network dynamics. The research team further explains how the operations of AM-ART replicate the identified memory management procedures in the base model. Moreover, they propose how to apply AM-ART to quantitatively study memory loss in general population, aged population and dementia patients. With the help of the parameterized computational autobiographical memory model, they can evaluate the phenomena of memory loss in a rapid and quantitative manner, which may be difficult or impossible in human participants.

Going forward, the research team aims to investigate the roles of autobiographical memory in more complex cognitive tasks and will look into the interactions between the autobiographical memory module and other memory modules in human brain.

Citation:
Di Wang, Ah Hwee Tan, Chunyan Miao, and Ahmed Moustafa. “Neurocomputational modeling of memory loss”. In Proceedings of International Conference on Ageless Ageing, Beijing, China. 2017.

Back to listing