Modeling memory and perception
Section snippets
Overview
Events and memory traces are represented as vectors of feature values. The traces are stored separately, although some are formed through accumulation of multiple events. Traces of new events are stored in memory as incomplete and error prone vectors. Such traces are termed episodic, and contain context, physical, and meaning features. When an event is similar to one or more earlier events (e.g., a nominal repetition is taking place) there is a tendency to store a new episodic trace, but also a
Recognition tasks
In Shiffrin and Steyvers (1997), we made a number of simplifying assumptions enabling Bayesian analysis to be applied to single-item episodic recognition—i.e., presentation of a list of words followed by a list of test words, half of which were on the list and half new (a very similar episodic recognition model was developed in parallel by McClelland & Chappell, 1998). It was assumed that all study and test words were randomly generated vectors chosen according to a geometric distribution with
Retrieval of knowledge and priming
The REM system can and ought to be applied to the retrieval of knowledge: A memory framework can hardly be complete without modeling knowledge retrieval, both in its own right, and because such retrieval is critical to explain episodic storage and retrieval. By knowledge we refer not only to facts, and verbal knowledge, but perception generally, motor programs, and everything else that is learned over development. Furthermore, our view is that such learning lies on a continuum: Although we have
Extensions
The studies and modeling described in this article, of course, represent only a current and personal snapshot of a continuing journey of research that is just beginning and will not near completion for many lifetimes. I hope and believe this theory represents systematic progress that builds on the great mass of collected data and many earlier theories, including my own (e.g., Atkinson & Shiffrin, 1968, Shiffrin, 1970, Raaijmakers & Shiffrin, 1980, Raaijmakers & Shiffrin, 1981, Gillund &
Final note
This article was meant to summarize the present status of a long-term research project through a conceptual overview. Space limitations required omissions such as almost all formal analysis, even though the article’s genesis was the award of the David E. Rumelhart Prize for contributions to the ‘the formal analysis of human cognition.’ I want to assure the reader that formal analysis is available in the typically much longer publications cited. The present overview emphasizes the role of
Acknowledgments
This research was supported by NIMH Grants 12717 and 63996, and numerous students, postdoctoral visitors, and research collaborators.
References (52)
- et al.
Relations between prototype, exemplar, and decision bound models of categorization
Journal of Mathematical Psychology
(1993) - et al.
Midazolam amnesia and dual-process models of the word frequency mirror effect
Journal of Memory and Language
(2002) - Anderson, J. (1990). The adaptive character of thought. Hillsdale, NJ:...
The adaptive nature of human categorization
Psychological Review
(1991)- et al.
Toward a unified theory of similarity and recognition
Psychological Review
(1988) - Atkinson, R. C., & Shiffrin, R. M. (1968). Human memory: A proposed system and its control processes. In K. W. Spence &...
Priming is not all bias: Commentary on Ratcliff and McKoon (1997)
Psychological Review
(1999)- Burgess, C., & Lund, K. (2000). The dynamics of meaning in memory. In E. Dietrich & A. B. Markman (Eds.), Cognitive...
- Criss, A. H., & Shiffrin, R. M. (submitted, a). Context noise and item noise jointly determine recognition memory: A...
- Criss, A. H., & Shiffrin, R. M. (submitted, b). The lack of interference for pairs and single items in episodic memory....
A context noise model of episodic word recognition
Psychological Review
An ARC-REM model for accuracy and response time in recognition and recall
Journal of Experimental Psychology: Learning, Memory, and Cognition
Illusions, perception, and Bayes
Nature Neuroscience
A retrieval model for both recognition and recall
Psychological Review
The mirror effect in recognition memory: Data and theory
Journal of Experimental Psychology: Learning, Memory, and Cognition
Judgments of frequency and recognition memory in a multiple trace memory model
Psychological Review
Effects of similarity and repetition on memory: Registration without learning?
Journal of Experimental Psychology: Learning, Memory, and Cognition
Mechanisms of source confusion and discounting in short-term priming 2: Effects of prime similarity and target duration
Journal of Experimental Psychology: Learning, Memory, & Cognition
Perception and preference in short-term word priming
Psychological Review
Mechanisms of source confusion and discounting in short-term priming. 1: Effects of prime duration and prime recognition
Memory & Cognition
A solution to Plato’s problem: The Latent Semantic Analysis theory of acquisition, induction, and representation of knowledge
Psychological Review
Feature frequency effects in recognition memory
Memory & Cognition
Cited by (52)
Rational use of episodic and working memory: A normative account of prospective memory
2021, NeuropsychologiaCitation Excerpt :The model contains no explicit retrieval processes; rather, the model conditions its responses on all the information in the noisy memory store. This global-parallel property is consistent with most mathematical models of memory retrieval that assume parallel contact with all memory elements (e.g., Shiffrin, 2003). But the model provides accounts of reaction times and error rates without assuming a retrieval process that yields an item or subset of items from memory that must then be further processed.
Deep memory distortions
2021, Cognitive PsychologyPreparation and measurement in quantum memory models
2018, Journal of Mathematical PsychologyCitation Excerpt :None of the models we use in psychology or cognitive science, at least for any behavioral tasks I find to be of any interest, are correct. We build models to increase our understanding of, and to slightly better approximate, the incredibly complex cognitive systems that determine behavior. (Shiffrin, 2003) However, this pragmatism raises an interesting point.
Modeling graded response congruency effects in task switching
2014, Acta PsychologicaCitation Excerpt :Unfortunately, many of these endeavors have been one-off efforts, with the models not investigated beyond the original articles in which they were proposed. Consequently, there has been little in the way of cumulative model development in task switching compared with other domains (e.g., Anderson, 2007; Logan, 2004; Perry, Ziegler, & Zorzi, 2007; Shiffrin, 2003). An exception is the model of compound cue retrieval proposed by Schneider and Logan (2005), which accounts for how responses are selected in task-switching situations.
Reciprocal relations between cognitive neuroscience and formal cognitive models: Opposites attract?
2011, Trends in Cognitive SciencesA memory-based model of Hick's law
2011, Cognitive PsychologyCitation Excerpt :These two points highlight a valuable approach to cognitive modeling: cumulative model development. The essence of the approach is that new models are built on existing models, inheriting their successes while extending them to explain more phenomena (e.g., Anderson, 2007; Logan, 2002, 2004; Perry, Ziegler, & Zorzi, 2007; Shiffrin, 2003). The modeling work presented in this article is an example of the approach: We took the well-established mechanism for declarative memory retrieval in ACT-R and supplemented it with a mechanism for producing repetition effects to create a memory-based model that accounts for the main findings related to Hick’s law.