Modelling the Decay of Declarative Knowledge Between Learning Sessions

Boie, Felix (2020) Modelling the Decay of Declarative Knowledge Between Learning Sessions. Master's Thesis / Essay, Human-Machine Communication.

Preview	Text mHMC_2020_BoieF.pdf Download (4MB) | Preview
	Text toestemming.pdf Restricted to Registered users only Download (121kB)

Abstract

Adaptive learning systems adapt to the individual learner to optimize learning outcomes. By modelling the strength of facts in human memory over time, an adaptive system can present facts at the optimal time, right before they are forgotten. Prior studies have shown that facts decay more slowly between learning sessions than within learning sessions. In this thesis, we confirm this finding in naturalistic learning data collected using an adaptive learning system in two university courses. We demonstrate that while the system’s current ACT-R memory model captures within-session memory performance well, it does not adequately capture between-session memory decay. To better account for this, we extend the model by scaling the time between sessions by a psychological time factor (PTF). Here we show in detail how the PTF improves the learning system. Our findings suggest that the optimal PTF depends on the interval between sessions and is affected by sleep. Specifically, the PTF decreases as sessions are spaced further apart and remains constant once learners are thought to have slept between sessions. This thesis demonstrates the need of accounting for the passage of time in a more refined way than just using a single scaling factor for the PTF.

Item Type:	Thesis (Master's Thesis / Essay)
Supervisor name:	Borst, J.P. and Velde, M.A. van der and Rijn, D.H. van
Degree programme:	Human-Machine Communication
Thesis type:	Master's Thesis / Essay
Language:	English
Date Deposited:	12 Jan 2021 09:59
Last Modified:	12 Jan 2021 09:59
URI:	https://fse.studenttheses.ub.rug.nl/id/eprint/23792

Actions (login required)

View Item