Topic
Memory augmentation
About: Memory augmentation is a research topic. Over the lifetime, 57 publications have been published within this topic receiving 543 citations.
Papers published on a yearly basis
Papers
1 Jan 2002
TL;DR: Results show that memory for important computing events is fragile and that software tools could be used to augment users’ memories of how they have spent their time while computing.
Abstract: In pursuit of computational tools for augmenting computer users’ abilities to interleave multiple tasks, we examined computer users’ ability to identify and recall computing events deemed to be important, both with and without supportive reminder tools. Memory for events occurring during computer sessions was studied both 24 hours after an initial taped session and again after a one-month period of time. Results show that memory for important computing events is fragile and that software tools could be used to augment users’ memories of how they have spent their time while computing. In addition, we observed that approximately half of the events that users identified as important could be identified automatically with available computational methods, and an attempt was made to characterise the nature of the remaining events. Finally, in a probe of alternate designs for reminding systems, we found that users typically preferred to see snapshots of their computing events in a prototype reminder system, without audio, as opposed to a full video version of an event reminder system.
89 citations
TL;DR: This paper describes a Microsoft research initiative that focuses on memory augmentation, primarily based on the MyLifeBits Project of Microsoft researcher, Gordon Bell, who collected real life data using sensors and a still camera, called SenseCam, with the aim of developing a improved means of information retrieval.
Abstract: This paper describes a Microsoft research initiative that focuses on memory augmentation. The research is primarily based on the MyLifeBits Project of Microsoft researcher, Gordon Bell, who, for two years, collected real life data using sensors and a still camera, called SenseCam, with the aim of developing a improved means of information retrieval. With the MyLifeBits software, the SenseCam, and the low cost of memory of all kinds, the rebirth of the PC into a personal mainframe can be realized.
83 citations
TL;DR: In this paper, a memory-augmented proximal mapping module (MAPMM) was proposed by combining two types of memory augmentation mechanisms, namely high-throughput short-term memory (HSM) and cross-stage long-term Memory (CLM), which greatly reduced information loss between adjacent stages.
Abstract: Mapping a truncated optimization method into a deep neural network, deep
unfolding network (DUN) has attracted growing attention in compressive sensing
(CS) due to its good interpretability and high performance. Each stage in DUNs
corresponds to one iteration in optimization. By understanding DUNs from the
perspective of the human brain's memory processing, we find there exists two
issues in existing DUNs. One is the information between every two adjacent
stages, which can be regarded as short-term memory, is usually lost seriously.
The other is no explicit mechanism to ensure that the previous stages affect
the current stage, which means memory is easily forgotten. To solve these
issues, in this paper, a novel DUN with persistent memory for CS is proposed,
dubbed Memory-Augmented Deep Unfolding Network (MADUN). We design a
memory-augmented proximal mapping module (MAPMM) by combining two types of
memory augmentation mechanisms, namely High-throughput Short-term Memory (HSM)
and Cross-stage Long-term Memory (CLM). HSM is exploited to allow DUNs to
transmit multi-channel short-term memory, which greatly reduces information
loss between adjacent stages. CLM is utilized to develop the dependency of deep
information across cascading stages, which greatly enhances network
representation capability. Extensive CS experiments on natural and MR images
show that with the strong ability to maintain and balance information our MADUN
outperforms existing state-of-the-art methods by a large margin. The source
code is available at https://github.com/jianzhangcs/MADUN/.
82 citations
17 Oct 2021
TL;DR: Memory-Augmented Deep Unfolding Network (MADUN) as discussed by the authors proposes a memory-augmented proximal mapping module (MAPMM) by combining two types of memory augmentation mechanisms, namely High-throughput Short-term Memory (HSM) and Cross-stage Long-Term Memory (CLM), which greatly reduces information loss between adjacent stages.
Abstract: Mapping a truncated optimization method into a deep neural network, deep unfolding network (DUN) has attracted growing attention in compressive sensing (CS) due to its good interpretability and high performance. Each stage in DUNs corresponds to one iteration in optimization. By understanding DUNs from the perspective of the human brain's memory processing, we find there exists two issues in existing DUNs. One is the information between every two adjacent stages, which can be regarded as short-term memory, is usually lost seriously. The other is no explicit mechanism to ensure that the previous stages affect the current stage, which means memory is easily forgotten. To solve these issues, in this paper, a novel DUN with persistent memory for CS is proposed, dubbed Memory-Augmented Deep Unfolding Network (MADUN). We design a memory-augmented proximal mapping module (MAPMM) by combining two types of memory augmentation mechanisms, namely High-throughput Short-term Memory (HSM) and Cross-stage Long-term Memory (CLM). HSM is exploited to allow DUNs to transmit multi-channel short-term memory, which greatly reduces information loss between adjacent stages. CLM is utilized to develop the dependency of deep information across cascading stages, which greatly enhances network representation capability. Extensive CS experiments on natural and MR images show that with the strong ability to maintain and balance information our MADUN outperforms existing state-of-the-art methods by a large margin. The source code is available at https://github.com/jianzhangcs/MADUN/.
48 citations
7 May 2016
TL;DR: A five-week user study explores both individual requirements for video summaries and the differences in cognitive load, user experience, memory experience, and recall experience between review using video summarisations and non-summary review techniques to inform the design of future lifelogging data summarisation systems for memory augmentation.
Abstract: Reviewing lifelogging data has been proposed as a useful tool to support human memory. However, the sheer volume of data (particularly images) that can be captured by modern lifelogging systems makes the selection and presentation of material for review a challenging task. We present the results of a five-week user study involving 16 participants and over 69,000 images that explores both individual requirements for video summaries and the differences in cognitive load, user experience, memory experience, and recall experience between review using video summarisations and non-summary review techniques. Our results can be used to inform the design of future lifelogging data summarisation systems for memory augmentation.
47 citations
Related Topics (5)
Performance Metrics
| Year | Papers |
|---|---|
| 2021 | 8 |
| 2020 | 3 |
| 2019 | 10 |
| 2018 | 6 |
| 2017 | 3 |
| 2016 | 5 |