Optimal forgetting: Semantic compression of episodic memories.

David G Nagy,Balázs Török,Gergő Orbán,Samuel J Gershman

doi:10.1371/journal.pcbi.1008367

Abstract

It has extensively been documented that human memory exhibits a wide range of systematic distortions, which have been associated with resource constraints. Resource constraints on memory can be formalised in the normative framework of lossy compression, however traditional lossy compression algorithms result in qualitatively different distortions to those found in experiments with humans. We argue that the form of distortions is characteristic of relying on a generative model adapted to the environment for compression. We show that this semantic compression framework can provide a unifying explanation of a wide variety of memory phenomena. We harness recent advances in learning deep generative models, that yield powerful tools to approximate generative models of complex data. We use three datasets, chess games, natural text, and hand-drawn sketches, to demonstrate the effects of semantic compression on memory performance. Our model accounts for memory distortions related to domain expertise, gist-based distortions, contextual effects, and delayed recall.

Highlights

It has long been known that human memory is far from an exact reinstatement of past sensory experience
We address two challenges. 1, The environmental statistics is not known for the brain, rather these have to be learned over time from limited observations. 2, Information theory does not specify how different distortions of original experiences should be penalised
We show that compression of experiences through a generative model gives rise to systematic distortions that qualitatively correspond to a diverse range of observations in the experimental literature

Summary

Introduction

It has long been known that human memory is far from an exact reinstatement of past sensory experience. A great number of experiments have shed light on systematic ways in which the distortions in recalled memories can be influenced both by past and future information, as well as the context of encoding and recall. Canonical examples of past knowledge influencing recall include experiments of Bartlett [4] where for folk tales recalled by subjects of non-matching cultural background, the recalled versions were found to be modified in ways that made the stories more consistent with the subjects’ cultural background, leading to the suggestion that memory seems to be more reconstructive than reproductive. Paradigmatic examples of memory disruptions due to information obtained after the experience being recalled include post-event misinformation [8], imagination inflation [9], hindsight bias [10], or leading questions [11]. The rich set of systematic distortions provides insights into the principles governing memory formation, which provides a means to predict how experiences are transformed in memory

Methods

Results

Discussion

Conclusion