External Memory Research Articles

Some Memory Devices of Early Chinese Textile Technologies and Artisans’ Transmission of Knowledge

Abstract Patterned silks served as signifiers for rituals and status in early China. They were patterned using different techniques, for example, some sha-silks from Mawangdui dated to 168 CE were repeatedly stamped with motifs while other silks were woven with monochrome or polychrome motifs. The technical knowledge for producing such silks was embodied as internal memory in the artisans who predominantly worked in state- run workshops until the mid-first century BCE. Artisans also invented external memory devices to facilitate their work. Some devices – bronze stamps and pattern-shafts on looms – have been unearthed from Guangzhou and Laoguanshan dated to 122 and 118 BCE respectively. Most techniques did not circulate widely, and over time, some disappeared and were then replaced by others. Lacking artisans’ own accounts, this article examines these external memory devices in the workshop context to suggest the hypothesis that the transmission of the embodied knowledge depended on the mobility and freedom of the artisans.

Ars Memoriae

Ars Memoriae, The Art to Remember analyzes the role of art within the larger history and evolution of external memory devices. It looks at material traces of remembering and the invention of an ever-changing body of language expressions, like signs and symbols, to enhance communication capabilities. I followed the process of externalizing emotions, knowledge, and information, starting in the Palaeolithic Stone Age about 3 million years ago, until, in a speculative future, it will be internalized again using artificial wetware, neuro-computers, and DNA coding. > Click on the image to download the PDF.

Enhancing IoT data acquisition efficiency via FPGA-based implementation with OpenCL framework

The increasing demand for real-time data processing in Internet of Things (IoT) applications necessitates the development of efficient and flexible data acquisition systems capable of receiving and processing data from various sensor types. In conjunction with OpenCL, field-programmable gate arrays (FPGAs) have recently emerged as powerful platforms for accelerating data-intensive tasks. This study explored the implementation of an FPGA for data acquisition using OpenCL, aiming to design and implement an efficient data acquisition system tailored for IoT applications. Utilizing OpenCL for FPGA-based data acquisition offers several advantages that contribute to system efficiency, particularly in hardware interfaces between FPGA and external devices used in IoT applications. OpenCL abstracts the complexity of the FPGA hardware interface to external DDR memory for storing temporary data and a communication interface to the host CPU for transferring the collected data and enabling remote access, enabling developers to focus on algorithm design and functionality. To enable data reading from an external analog-to-digital converter (ADC) chip for IoT applications, we developed a component module that utilizes the Avalon-streaming interface and can stream the data to the OpenCL kernel. An experiment was conducted to demonstrate the performance of our proposed design. According to the findings of the experiments, a data acquisition implementation based on an FPGA and OpenCL can simultaneously read analog signals via a multichannel ADC. The proposed design provides a foundation for designing efficient data acquisition solutions, addressing the increasing needs of FPGA-based data acquisition in various IoT environments.

Comparative analysis of remote user authentication schemes based on external memory

Because of recent availability of remote services and resources, remote user authentication becomes an essential component of all digital environment. Remote user login authentication is the process of validating the identity of a user. Users present their credentials, such as username and password, as evidence of their identity. The most practical and effectively implemented remote user authentication scheme is the smart card-based one, however, because it is expensive to acquire and operate smart card facilities, users may find it challenging to employ smart card authentication schemes in remote environments. Mobility and simplicity of acquisition—especially for remote access—are the benefits of using this external memory. Therefore, this research performs comparative analysis of these remote user authentication schemes based external memory. The schemes are first evaluated for security features and performance in terms of computation cost. The results of analysis shown that Buhari et al.’s scheme has the highest security feature and is the only scheme that uses light-weight tamper resistance client file, followed the reviewed Kumari et al.’s scheme which is the only scheme that handles user privacy. The most efficient scheme is Buhari et al.’s scheme followed by Rhee et al.’s scheme.

Mitigating the Impact of Lognormal Atmospheric Turbulence Channel Estimation on FSO Communication Systems Using Advanced Deep Learning Modules

ABSTRACTCurrently, researchers and commercial entities worldwide are highly interested in optical wireless communication links operating in terrestrial environments. These links offer the distinct advantage of enabling high‐speed data transmission while keeping operational and installation costs low, all without requiring licensing. Free‐space optical (FSO) communication has established itself as a reliable method for delivering ultra‐high data‐rate services. However, the presence of atmospheric turbulence‐induced fading poses a significant challenge for FSO communication systems, particularly when operating over channels affected by time‐varying turbulence. The effectiveness of FSO systems is greatly influenced by atmospheric conditions in a specific area because the laser beam propagates through the atmosphere. As a consequence, this can lead to a significant degradation such as high complexity, high estimation error, and high bit error rate (BER) in the performance of FSO systems. This paper presents a stacked dual attention LSTM network with layered neural Turing machine (SDALSTM‐LaNTM) for FSO channel estimation over lognormal turbulence‐induced fading channels. In this approach, the stacked structure ensures a robust feature extraction capacity of SDALSTM and enhances its ability to capture intricate relationships among multiple atmospheric variables. After employing the stacked LSTM networks, the extracted features are combined with the augmented features obtained through the dual attention (DAttn) mechanism by concatenation. Then the LaNTM architecture used external memory to solve the channel parameter estimation problem and update new set of hidden features for estimation task. The ability to access and update memory enables the model to capture long‐term dependencies and context, making it beneficial for channel prediction in FSO systems. Presently, we are in the process of verifying the channel estimation model using measured FSO channel data. Simulation results clearly demonstrate the outstanding estimation performance of the SDALSTM‐LaNTM model in FSO systems, specifically when dealing with lognormal turbulence‐induced fading channels. Moreover, we highlight the effectiveness of the SDALSTM‐LaNTM model by conducting a comprehensive comparison with existing models, considering factors such as average capacity performance, outage probability estimation, and BER.

The group mind of hybrid teams with humans and intelligent agents in knowledge-intense work

Studies regularly demonstrate how well intelligent agents (IAs) can support humans or are demonstrably superior to them in some areas. Given that some tasks likely remain unsuitable for even the most intelligent machines in the mid-future, work in hybrid teams of humans and IAs—where the capabilities of both are effectively combined—will most likely shape the way we work in the coming decades. In an abductive study, we investigate an early example of hybrid teams, consisting of a conversational intelligent agent (IA) and humans, that aims to improve health behavior or change personality traits. We theorize Transactive Intelligent Memory System (TIMS) as a new vision of collaboration between humans and IAs in hybrid teams, based on our empirical insights and our literature review on transactive memory systems theory. Our empirical evidence shows that IAs can develop a form of individual and external memory, and hybrid teams of humans and IAs can realize joint systems of transactive memory—a competence that current literature only ascribes to humans. We further find that whether individuals view IAs merely as external memory aids or as part of their teams’ transactive memory is moderated by the tasks’ complexity and knowledge intensity, as well as the IA's ability to complete the task. This theorizing helps to better understand the role of IAs in future team-based working processes. Developers of IAs can use TIMS as a tool for requirements formulation to prepare their software agents for collaboration in hybrid teams.

Maximizing the "Functional" in the Functional External Memory Aid Tool for Patient-Centered Treatment Planning in Mild Cognitive Impairment.

The Functional External Memory Aid Tool (FEMAT) is an 11-item performance-based measure that simulates everyday tasks (e.g., medication management) to measure one's use of compensation (e.g., pill organizer). The FEMAT is not a norm-referenced diagnostic tool; rather, it is designed to provide clinicians with information about a patient's function and compensation in a standardized format. To provide further evidence for validity of the FEMAT, the purpose of this study was to conduct a mixed-methods analysis of FEMAT responses to operationalize the types of behaviors elicited during test administration. We adopted an embedded single-case study design to analyze the FEMAT administrations of 12 community-dwelling female participants, ages 73-90 years, who met criteria for mild cognitive impairment. These participants were part of the original validation sample during which they completed the FEMAT (Version 1.0) in a single session with an assessor. Sessions were video-recorded and transcribed. We analyzed the data qualitatively in two phases using theme and subtheme codes and then quantitatively analyzed the data for response trends. Our thematic analysis revealed four themes to operationalize participants' verbal and behavioral FEMAT responses: (a) used an external memory aid (EMA), (b) described using a strategy, (c) discussed memory and comprehension, and (d) discussed functional information. While completing the FEMAT, most participants described using a strategy (85%) and discussed their self-perceived memory and comprehension (33%). Several subthemes also emerged within each theme (e.g., self-perceived memory strengths and barrier). Our analysis suggests that beyond a 4-point item score, the FEMAT elicits behavioral (e.g., used an EMA) and verbal (e.g., discussion of health information) content that may be useful for person-centered treatment planning. This study is part of a comprehensive research agenda establishing the evidence for the validity of the FEMAT to support its use to inform person-centered treatment approaches.

Enhancing Graph Neural Networks via Memorized Global Information

Graph neural networks (GNNs) have gained significant attention for their impressive results on different graph-based tasks. The essential mechanism of GNNs is the message-passing framework, whereby node representations are aggregated from local neighborhoods. Recently, Transformer-based GNNs have been introduced to learn the long-range dependencies, enhancing performance. However, their quadratic computational complexity, due to the attention computation, has constrained their applicability on large-scale graphs. To address this issue, we propose MGIGNN ( M emorized G lobal I nformation G raph N eural N etwork), an innovative approach that leverages memorized global information to enhance existing GNNs in both transductive and inductive scenarios. Specifically, MGIGNN captures long-range dependencies by identifying and incorporating global similar nodes, which are defined as nodes exhibiting similar features, structural patterns and label information within a graph. To alleviate the computational overhead associated with computing embeddings for all nodes, we introduce an external memory module to facilitate the retrieval of embeddings and optimize performance on large graphs. To enhance the memory-efficiency, MGIGNN selectively retrieves global similar nodes from a small set of candidate nodes. These candidate nodes are selected from the training nodes based on a sparse node selection distribution with a Dirichlet prior. This selecting approach not only reduces the memory size required but also ensures efficient utilization of computational resources. Through comprehensive experiments conducted on ten widely-used and real-world datasets, including seven homogeneous datasets and three heterogeneous datasets, we demonstrate that our MGIGNN can generally improve the performance of existing GNNs on node classification tasks under both inductive and transductive settings.

All-optical organic photochemical integrated nanophotonic memory: low-loss, continuously tunable, non-volatile

Controlling changes in the optical properties of photonic devices allows photonic integrated circuits (PICs) to perform useful functions, leading to a large breadth of applications in communications, computing, and sensing. Many mechanisms to change optical properties exist, but few allow doing so in a reversible, non-volatile manner. Without such mechanisms, power inefficiencies and use of external memory are inevitable. In this work, we propose and experimentally demonstrate reversible, non-volatile phase actuation of a silicon nitride PIC with thermally stable photochromic organic molecules vapor-deposited within a slot waveguide structure. The use of a high-core-index platform allows the photochemical phase actuation of a planar-resonator-based photonic memory unit, which enables positive and negative signal weighting and permits integrated spectroscopic analysis. We show properties of this all-optical memory for a silicon photonics platform, including low loss in the optical C-band, first-order photokinetics of the photoconversion, bidirectional scalable switching, and continuous tuning. Such features are critical for memories in analog applications such as quantum, microwave, and neuromorphic photonics, where bipolar weights, low loss, and precision are paramount. More generally, this work suggests that back-end-of-line-compatible vapor deposition of organic molecules into silicon photonic circuits is promising to introduce non-silicon-native functionality.

The invariant measure of a walking droplet in hydrodynamic pilot–wave theory

We study the long time statistics of a walker in a hydrodynamic pilot-wave system, which is a stochastic Langevin dynamics with an external potential and memory kernel. While prior experiments and numerical simulations have indicated that the system may reach a statistically steady state, its long-time behavior has not been studied rigorously. For a broad class of external potentials and pilot-wave forces, we construct the solutions as a dynamics evolving on suitable path spaces. Then, under the assumption that the pilot-wave force is dominated by the potential, we demonstrate that the walker possesses a unique statistical steady state. We conclude by presenting an example of such an invariant measure, as obtained from a numerical simulation of a walker in a harmonic potential.

Transcranial direct current stimulation over medial prefrontal cortex reduced alpha power and functional connectivity during somatic but not semantic self-referential processing

Past self-report and cognitive-behavioural studies of the effects of transcranial direct current stimulation (tDCS) targeting the medial prefrontal cortex (mPFC) on semantic self-referential processing (SRP) have yielded mixed results. Meanwhile, electroencephalography (EEG) studies show that alpha oscillation (8–12 Hz) may be involved during both semantic and somatic SRP, although the effect of tDCS on alpha-EEG during SRP remains unknown. The current study assessed the EEG and subjective effects of 2 mA tDCS over the mPFC while participants were SRP either on semantic (life roles, e.g., “friend”) or somatic (outer body, e.g., “arms”) self-referential stimuli compared to resting state and an external attention memory task in 52 young adults. Results showed that whereas mPFC-tDCS did not yield significant changes in participants’ mood or experienced attention or pleasantness levels during the SRP task, EEG source analysis indicated, compared to sham stimulation, that tDCS reduced alpha power during somatic but not semantic SRP in the posterior cingulate cortex (PCC), and the frontal, parietal, temporal, and somatosensory cortex, and reduced the functional connectivity between the left inferior parietal lobule and the ventral PCC, but only when mPFC-tDCS was applied at the second while not the first experimental session. Our results suggest that while mPFC-tDCS may be insufficient to alter immediate subjective experience during SRP, mPFC-tDCS may modulate the power and functional connectivity of the brain’s alpha oscillations during somatic SRP. Future research directions are discussed.

Learning in the wild – exploring interactive notifications to foster organic retention of everyday media content

ABSTRACT The amount of information we consume daily through smartphones has prompted the development of Personal Knowledge Management Systems (PKMS). However, these external memory aids can lead to overreliance and do not help to keep information in our organic memory. This paper presents the design and evaluation of MemoryMate, an app that fosters recall of saved content through interactive notifications. The app's design is based on insights from a focus group (N = 7) that inquired about people's expectations and preferences regarding the notification design. The results of a three-week in-the-wild study (N = 23) of MemoryMate suggest that notifications could trigger engagement with saved content, and two-thirds of participants felt notifications improved their memory and retention. Retrieval practice notifications with interactive tasks performed well and received positive feedback. Content priority and time since the last interaction influenced notification engagement, emphasising the need for periodic reminders to maintain user interest in saved content. We discuss potential implications for the design of PKMSs as a tool for lifelong learning.

Towards Retrieval-Augmented Architectures for Image Captioning

The objective of image captioning models is to bridge the gap between the visual and linguistic modalities by generating natural language descriptions that accurately reflect the content of input images. In recent years, researchers have leveraged deep learning-based models and made advances in the extraction of visual features and the design of multimodal connections to tackle this task. This work presents a novel approach toward developing image captioning models that utilize an external k NN memory to improve the generation process. Specifically, we propose two model variants that incorporate a knowledge retriever component that is based on visual similarities, a differentiable encoder to represent input images, and a k NN-augmented language model to predict tokens based on contextual cues and text retrieved from the external memory. We experimentally validate our approach on COCO and nocaps datasets and demonstrate that incorporating an explicit external memory can significantly enhance the quality of captions, especially with a larger retrieval corpus. This work provides valuable insights into retrieval-augmented captioning models and opens up new avenues for improving image captioning at a larger scale.

Development of a soil electrical conductivity measurement system in paddy fields

This study aims to develop a data collection system for measuring soil electrical conductivity (EC) in Paddy Fields using the Wenner measurement principle, with one electrode configuration based on the Veris 3100 measuring machine, and a second electrode configuration with fixed electrodes. The result is a system that measures soil EC using the fixed electrode configuration for more stable results compared to the coulter electrode configuration on the Veris 3100. This system comprises three main components: a 7-inch industrial touchscreen monitor used for real-time monitoring and storing collected data on external memory, a controller that provides reverse direct current (DC) power supply to the electrodes and measures voltage and current parameters on the electrodes, two mechanical configurations including six electrodes in contact with the soil. The system is developed to operate best after soil preparation and before seed sowing. The collected data is processed and compared to actual measurements using an artificial neural network (ANN), resulting in an R2=0.7449, equivalent to linear regression. The system has been successfully installed and operated on a handheld tractor and tested in Paddy Fields on the outskirts of Can Tho City, Vietnam.

P‐57: A Novel Compression Algorithm using Machine Learning for Mura Compensation of OLED Panel

In the modern display panel manufacturing industry, the Mura problem that occurs during initial shipment has a significant impact on improving image quality. The purpose of this paper is to effectively solve these Mura defects while reducing additional production costs. We propose an innovative method of storing and compressing the De‐Mura compensation data of the display panel in external memory. Two‐dimensional compensation data compression based on machine learning using surface fitting and vector quantization is presented as an effective response to the Mura problem. This approach, which achieves up to 8:1 compression ratio while minimizing image quality degradation through vector quantization, is expected to provide economic benefits to manufacturers. The paper also introduces a new surface prediction method that replaces the existing DC‐shift and uses residual vector quantization, which can reduce the implementation complexity of vector quantization. Through experimental results, this study demonstrates that the compensated data compression technique yields excellent results in post‐compression restoration performance.

Hardware Efficient Integrated In-loop Filter for HEVC Encoder

The deblocking filter (DF) and the sample adaptive offset (SAO) filter, which aids in enhancing the subjective quality of the image, make up the in-loop filter of the high-efficiency video coding (HEVC) encoder and decoder. The in-loop filter significantly increases the computational load on the HEVC encoder. It is challenging to design an in-loop filter on hardware that can handle intensive computations while using the least amount of on-chip memory, taking external memory traffic and dependencies simultaneously delivering high throughput to support Ultra HD video applications. The proposed design employs the following strategies to address these issues. This work proposes an address generation technique for pipelined horizontal and vertical filtering in DF, that avoids a transpose buffer which otherwise is required. This enables easy pipelining and parallelization thus improving throughput while reducing the on-chip memory utilization. A simplified SAO filter with parallel-pipelined processing is included in the design. These features enable the design to support ultra-HD 7680 × 4320 @ 40 fps video applications. The proposed hardware architecture has a total gate count of 7.73 K LUTs and 2.8 K slice registers, and it is implemented on a 28 nm field programmable gate array (FPGA) platform.

Architectural Support for Sharing, Isolating and Virtualizing FPGA Resources

FPGAs are increasingly popular in cloud environments for their ability to offer on-demand acceleration and improved compute efficiency. Providers would like to increase utilization, by multiplexing customers on a single device, similar to how processing cores and memory are shared. Nonetheless, multi-tenancy still faces major architectural limitations including: (a) inefficient sharing of memory interfaces across hardware tasks (HT) exacerbated by technological limitations and peculiarities, (b) insufficient solutions for performance and data isolation and high quality of service, and (c) absent or simplistic allocation strategies to effectively distribute external FPGA memory across HT. This article presents a full-stack solution for enabling multi-tenancy on FPGAs. Specifically, our work proposes an intra-fpga virtualization layer to share FPGA interfaces and its resources across tenants. To achieve efficient inter-connectivity between virtual FPGAs (vFGPAs) and external interfaces, we employ a compact network-on-chip architecture to optimize resource utilization. Dedicated memory management units implement the concept of virtual memory in FPGAs, providing mechanisms to isolate the address space and enable memory protection. We also introduce a memory segmentation scheme to effectively allocate FPGA address space and enhance isolation through hardware-software support, while preserving the efficacy of memory transactions. We assess our solution on an Alveo U250 Data Center FPGA Card, employing 10 real-world benchmarks from the Rodinia and Rosetta suites. Our framework preserves the performance of HT from a non-virtualized environment, while enhancing the device aggregate throughput through resource sharing; up to 3.96x in isolated and up to 2.31x in highly congested settings, where an external interface is shared across four vFPGAs. Finally, our work ensures high-quality of service, with HT achieving up to 0.95x of their native performance, even when resource sharing introduces interference from other accelerators.

Age‐related advantage for recall of complex naturalistic information following cognitive offloading

AbstractPrevious work on cognitive offloading has found that young and older adults create and use external memory aids to boost performance on simple memory‐based tasks. To date, little work has investigated whether cognitive offloading can benefit performance when the memoranda are complex and naturalistic. In this study, 64 participants (32 young adults aged 18–26 years and 32 older adults aged 60+ years) studied facts about health conditions, then completed free recall and old‐new recognition tests under two conditions. In the cognitive offloading choice condition, participants could create notes during study to use on upcoming memory tests. In the internal memory condition, participants had to rely on internal memory alone. Both age groups benefited from using cognitive offloading. Notably, older adults outperformed young adults on the free recall test when given the opportunity to offload. These results suggest that cognitive offloading can help older adults overcome everyday memory challenges.

MoMENt: Marked Point Processes with Memory-Enhanced Neural Networks for User Activity Modeling

Marked temporal point process models (MTPPs) aim to model event sequences and event markers (associated features) in continuous time. These models have been applied to various application domains where capturing event dynamics in continuous time is beneficial, such as education systems, social networks, and recommender systems. However, current MTPPs suffer from two major limitations, i.e., inefficient representation of event dynamic’s influence on marker distribution and losing fine-grained representation of historical marker distributions in the modeling. Motivated by these limitations, we propose a novel model called M arked P o int Processes with M emory- E nhanced N eural Ne t works (MoMENt) that can capture the bidirectional interrelations between markers and event dynamics while providing fine-grained marker representations. Specifically, MoMENt is constructed of two concurrent networks: Recurrent Activity Updater (RAU) to capture model event dynamics and Memory-Enhanced Marker Updater (MEMU) to represent markers. Both RAU and MEMU components are designed to update each other at every step to model the bidirectional influence of markers and event dynamics. To obtain a fine-grained representation of maker distributions, MEMU is devised with external memories that model detailed marker-level features with latent component vectors. Our extensive experiments on six real-world user interaction datasets demonstrate that MoMENt can accurately represent users’ activity dynamics, boosting time, type, and marker predictions, as well as recommendation performance up to 76.5%, 65.6%, 77.2%, and 57.7%, respectively, compared to baseline approaches. Furthermore, our case studies show the effectiveness of MoMENt in providing meaningful and fine-grained interpretations of user-system relations over time, e.g., how user choices influence their future preferences in the recommendation domain.

Cultural Adaptation of the Functional External Memory Aid Tool for Use in the United Kingdom and Australia

Purpose: The Functional External Memory Aid Tool (FEMAT) is a performance-based measure of applied cognitive-linguistic abilities and attempted use of compensatory aids and strategies (e.g., using a calendar to recall an appointment) through completion of simulated everyday living tasks. The FEMAT was originally developed in the United States (US), and a need arose from clinicians to adapt the FEMAT for use in other countries. The aim of this study is to describe the cultural adaptation process that resulted in the United Kingdom (UK) and Australian (AU) versions of the FEMAT via international clinical research partnerships. Method: The FEMAT was adapted into UK and AU cultures using established guidelines for cultural and linguistic adaptation. The adaptation was completed by researchers and clinicians from the US, UK, and AU, and we followed a six-step process: (a) preparation, (b) forward translation, (c) back translation, (d) committee review, (e) field testing, and (f) finalization. Results: The FEMAT items and materials were successfully adapted from the source culture (i.e., English-US) to the target cultures (i.e., English-UK, English-AU). Through the adaptation process, the FEMAT items were determined to be conceptually appropriate for each culture and minimal modifications were made to the FEMAT materials (e.g., voicemail recordings recorded in target dialect). The finalized FEMAT-English-UK and the FEMAT-English-AU versions are available for free on the FEMAT website. Conclusions: Through successful international clinical research collaborations, the FEMAT was successfully adapted for use in other English-speaking cultures. This work provides guidance to streamline future FEMAT cultural adaptations. Supplemental Material: https://doi.org/10.23641/asha.25564461