Encoding (memory)

About: Encoding (memory) is a research topic. Over the lifetime, 7547 publications have been published within this topic receiving 120214 citations. The topic is also known as: memory encoding & encoding of memories.

Skilled memory and expertise: Mechanisms of exceptional performance.

Abstract: : Skilled memory theory describes how subjects can acquire exceptional memory skills and thereby develop long-term memory with performance characteristics comparable to those of short-term memory. The research reported in this paper has further investigated the detailed mechanisms of exceptional digit span and has explored the generalizability of skilled memory theory to account for the superior memory of memory experts and of other experts in their domains of expertise. Two new studies tested three principles of skilled memory in the domain of exceptional digit span. One study showed that encoding a four- digit number as a unit (e.g., coding 3526 as a running time for a race) enables even expert runners to reliably retrieve only the first two digits of the number. The other study demonstrated that in addition to encoding numbers as running times, subjects encoded other patterns and relations between digits. This study also monitored in detail the emergence of a retrieval structure as a function of practice. A review of studies of individuals with exceptional memory shows that skilled memory theory can account for all available evidence on exceptional memory. Furthermore, detailed analyses of the memory performance of an exceptional waiter and a chess master support the claim that skilled memory theory can account for the superior memory performance of experts in their domain of expertise.

Learning to detect violent videos using convolutional long short-term memory

Abstract: Developing a technique for the automatic analysis of surveillance videos in order to identify the presence of violence is of broad interest. In this work, we propose a deep neural network for the purpose of recognizing violent videos. A convolutional neural network is used to extract frame level features from a video. The frame level features are then aggregated using a variant of the long short term memory that uses convolutional gates. The convolutional neural network along with the convolutional long short term memory is capable of capturing localized spatio-temporal features which enables the analysis of local motion taking place in the video. We also propose to use adjacent frame differences as the input to the model thereby forcing it to encode the changes occurring in the video. The performance of the proposed feature extraction pipeline is evaluated on three standard benchmark datasets in terms of recognition accuracy. Comparison of the results obtained with the state of the art techniques revealed the promising capability of the proposed method in recognizing violent videos.

Category size effects in recall: The roles of relational and individual item information.

Abstract: Memory for events varies as a function of the number of events in a given class, but previous research from organization theory did not succeed in establishing a consistent function relating memory and category size. We suggest that prior research can be systematized within a framework of relational and individual item processing. Relational processing refers to the encoding of similarities among events, and individual item, processing refers to encoding of distinctive information for each event. Assuming the importance of both types of information for precise recall and that the type of information encoded will depend on category size and the subject's attention to relational or distinctive features, predictions are derived concerning the interaction of orienting activity and category size. The predicted interaction was obtained in two experiments that demonstrated that small categories are better recalled following relational processing, and large categories are; better recalled following individual item processing. Additional dependent measures (clustering, category recall, items per category recall, and cued recall) provided highly consistent converging evidence for the proposed theoretical analysis. The general conclusion is that theories of memory must explain the paradoxical fact of the simultaneous importance of both similarity and difference.

Divided attention and indirect memory tests.

Abstract: Attentional state during acquisition is an important determinant of performance on direct memory tests. In two experiments we investigated the effects of dividing attention during acquisition on conceptually driven and data-driven indirect memory tests. Subjects read a list of words with or without distraction. Memory for the words was later tested with an indirect memory test or a direct memory test that differed only in task instructions. In Experiment 1, the indirect test was categoryexemplar production (a conceptually driven task) and the direct test was category-cued recall. In Experiment 2, the indirect test was word-fragment completion (a data-driven task) and the direct test was word-fragment cued recall. Dividing attention at encoding decreased performance on both direct memory tests. Of the indirect tests, category-exemplar production but not word-fragment completion was affected. The results indicate that conceptually driven indirect memory tests, like direct memory tests, are affected by divided attention, whereas data-driven indirect tests are not. These results are interpreted within the transfer-appropriate processing framework.

Memory for multidimensional source information.

Abstract: Source memory may comprise recollection of multiple features of the encoding episode. To analyze the simultaneous representation and retrieval of those multiple features, a multinomial memory model is presented that measures memory for crossed dimensions of source information. The first experiment investigated the validity of the new model. The model showed an excellent statistical fit to empirical data, and the parameters of multidimensional source memory were sensitive to manipulations of source similarity on distinct dimensions. The second experiment used the model to test the hypothesis that source memory for individual context attributes is stochastically related in the case of conscious recollection but independent in the case of familiarity-based recognition judgments. The prediction was supported by the introduction of a "remember"-"know" distinction in a multidimensional source memory test.