Current Memory Research Articles

Ubunye: An MEC Orchestration Service Based on QoE, QoS, and Service Classification Using Machine Learning

The increasing adoption of Internet of Things devices has led to a significant demand for cloud services, where latency and bandwidth play a crucial role in shaping users’ perception of network service quality. However, the use of cloud services with the desired quality is not always available to all users. Furthermore, uneven network coverage in urban and rural areas has created “digital deserts”, which are characterized by a lack of connectivity resources, complicating access to cloud services. In this scenario, edge computing emerges as a promising alternative for service provision. Edge computing leverages data processing at or near the source where it is generated rather than sending it to the cloud for processing. It can lead to several advantages, such as reduced latency and lower bandwidth usage. This paper addresses the need to ensure consistent quality of experience (QoE) and quality of service (QoS) in dynamic network environments, particularly in remote regions with limited infrastructure. We propose an orchestration service called Ubunye, which operates at the network edge and selects the most appropriate edge node to fulfill a given application request while satisfying its quality requirements. Ubunye considers factors such as latency and available bandwidth when selecting a node to execute the requested service. It implements a service classification system based on machine learning (ML) techniques. The ideal edge node is chosen through a multi-faceted evaluation, which includes current CPU load, memory availability, and other relevant parameters. Experiment results show that Ubunye effectively orchestrates resources at the network edge, enhancing QoE and QoS for services that demand low latency and high bandwidth. Additionally, it showcases the ability to classify services and allocate resources under challenging network conditions.

Development and Validity of the Adult Reading History Questionnaire (ARHQ) for Chinese.

The Adult Reading History Questionnaire (ARHQ) is a popular instrument for screening dyslexia in adults. Yet, there is no evidence suggesting its applicability in Chinese. The present study aims to develop a Chinese Adult Reading History Questionnaire (Chinese-ARHQ) and to assess its validity among Chinese adults. A total of 2439 Chinese adults completed the questionnaire of Chinese-ARHQ with 2234 questionnaires valid for further analysis (Nmales = 825, Mage = 26.29 years, SDage = 9.12). Among these participants, 184 completed the questionnaire twice and 221 completed reading-related tests. Results showed that the Chinese-ARHQ had good internal consistency reliability (Cronbach's α = 0.88), and the test-retest reliability was 0.81. Three factors (current reading, dyslexia symptoms and memory) were extracted through exploratory factor analysis. The scores of Chinese-ARHQ and its factors were both related to reading and phonological skills, revealing the good criterion validity of the questionnaire. Receiver operating characteristic (ROC) analysis revealed that the optimal threshold for the Chinese-ARHQ was 0.36 with a sensitivity of 83.3% and a specificity of 75.9%. Additionally, adults screened by the Chinese-ARHQ (> 0.36) performed worse in reading and phonological awareness, further validating the Chinese-ARHQ as a reliable screening tool for Chinese dyslexia in adults.

Dual‐Mode Flexible Parylene‐C‐Based In‐Memory Tactile Sensory Device with CMOS Compatibility

AbstractThe artificial tactile memory system enables electronic skin (e‐skin) electronics to detect and record tactile sensations like natural skin, significantly expanding the applications of flexible electronics in wearable devices, human‐computer interaction, and healthcare. However, current tactile memory electronics still face challenges such as complex structures and high hardware costs, limiting further high‐density integration. Herein, an in‐memory tactile sensory (IMTS) device with a simple structure is experimentally demonstrated, which can achieve pressure information detecting, storage, and erasure in a single device. Due to the chemical stability of the parylene‐C and its compatibility with standard photolithography, the device can be manufactured on advanced CMOS process platforms, showing potential for high‐density integration. Additionally, the device demonstrates excellent mechanical properties, maintaining outstanding electrical performance in both flat and bending conditions. Furthermore, pressure distribution can be detected and recorded with the IMTS arrays, providing new avenues for developing high‐performance tactile memory systems in human‐computer interaction and wearable electronics.

Rapid source forgetting across modalities: A problem for working memory models.

Working memory is a cognitive system that enables the temporary retention (usually a few seconds) of a limited amount of information. However, recent evidence has posed challenges to the conventional understanding of working memory's persistence. Chen et al. (Psychological Science, 29(4), 645-655, 2018) demonstrated that participants can easily make judgments using a stimulus's identity but cannot recall from which source the information came (presented either as a written word or a color patch) just milliseconds earlier. This "Source Amnesia" carries substantial implications for working memory models but has yet to be explored within the realm of verbal information. We fill this gap by investigating the robustness and generalizability of this rapid forgetting phenomenon. We first replicate the observed effect within the visual domain (Experiment 1) and subsequently extend it to the verbal domain (Experiment 2). Finally, we test the idea that participants may instead encode a positional context (Experiment 3), in line with the Interference model (Oberauer & Lin, Psychological Review, 124(1), 21, 2017). Aligning with the work of Chen et al. (Psychological Science, 29(4), 645-655,2018), our results consistently reveal a pronounced tendency for rapid forgetting, for both visual and verbal information regardless of whether the information is elicited for recall by format or position cues. The theoretical implications of these findings for current memory models are discussed.

Clinic evaluation of cognitive impairment in post-COVID syndrome: Performance on legacy pen-and-paper and new digital cognitive tests

BackgroundCognitive impairment, colloquially termed “brain fog”, is one of the most prevalent manifestations of post-Covid syndrome and a major contributor to impaired daily function and reduced quality of life. However, despite the high numbers of affected individuals presenting to clinical services with cognitive impairment, little work has been undertaken to date on the suitability of current memory clinic tests for identifying the cognitive deficits in this new acquired cognitive disorder.The aim of this study was therefore to determine the performance of people with post-Covid syndrome presenting with cognitive impairment on the Addenbrooke's Cognitive Examination-III (ACE-III), a cognitive test used widely in memory clinics. A subset of individuals also underwent testing on a novel battery of short digital tests assessing attention, speed of information processing and executive function, representing the domains primarily implicated in post-Covid cognitive dysfunction. Methods102 individuals with post-Covid syndrome presenting with subjective cognitive complaints were recruited from a specialist cognitive long Covid clinic at University Hospitals Sussex NHS Trust. All participants completed self-report questionnaires on depression, anxiety, sleep and fatigue. Cognitive performance was assessed using the ACE-III, with 20 participants also being tested on the digital Long COVID Assessment Battery (LCCAB) (N = 20). ResultsThe overall sample had a mean ACE-III score of 91/100 (SD 6) with 15.7% (16/102) scoring at or below the cut-off score considered to represent objective cognitive impairment. Of the 20 individuals who also completed the LCCAB, 89.47% were impaired on at least one task, primarily in the domains of attention, executive function and processing speed. Cognitive performance was not associated with depression, anxiety, sleep quality or fatigue. ConclusionThe vast majority of individuals with post-Covid syndrome presenting with subjective cognitive complaints do not exhibit impaired performance on the ACE-III. This likely reflects the historical use of ACE-III and other pen and paper cognitive tests to detect cognitive impairment in diseases causing dementia, but they are ill-equipped to identify impairment in those cognitive domains affected in post-Covid syndrome. The LCCAB detected cognitive impairments in nearly 90% of participants, primarily affecting attention, executive function, and processing speed. These observations highlight the need for alternative cognitive tests for use in routine clinical practice to detect the impairments in new acquired cognitive disorders that are not adequately captured by legacy tests.

ShieldCXL: A Practical Obliviousness Support with Sealed CXL Memory

The CXL (Compute Express Link) technology is an emerging memory interface with high-level commands. Recent studies applied the CXL memory expanding technique to mitigate the capacity limitation of the conventional DDRx memory. Unlike the prior studies to use the CXL memory as the capacity expander, this study proposes to use the CXL-based memory as a secure main memory device, while removing the conventional memory. In the conventional DDRx memory, to provide confidentiality, integrity, replay protection, and obliviousness, costly mechanisms such as counter-based integrity trees and location shuffling by ORAM (Oblivious RAM) are used. Such mechanisms incur significant performance degradation in the current DDR-based memory systems, and their costs increase as the capacity of the memory increases. To mitigate the performance degradation, the prior work proposed an obfuscated channel for a secure memory module enclosing its controller in the package. Based on the approach, we propose a secure CXL-only memory architecture called ShieldCXL . It uses the channel encryption and integrity protection mechanism of the CXL interface to provide a practical ORAM while supporting confidentiality, integrity, and replay protection from physical attacks and rowhammers. To protect the PCIe-connected memory expanding board, this study proposes to use the standard physical sealing technique to detect physical intrusion. To mitigate the increased latency with the sealed CXL memory module, the study further optimizes performance by adopting an in-package DRAM cache. In addition, this study investigates destination obfuscation when a CXL switch is used to route among multiple hosts and memory devices. The evaluation shows that ShieldCXL provides 9.16x performance improvements over the prior ORAM technique.

Metastable body-centered cubic CoMnFe alloy films with perpendicular magnetic anisotropy for spintronics memory

ABSTRACT A body-centered cubic (bcc) FeCo(B) is a current standard magnetic material for perpendicular magnetic tunnel junctions (p-MTJs) showing both large tunnel magnetoresistance (TMR) and high interfacial perpendicular magnetic anisotropy (PMA) when MgO is utilized as a barrier material of p-MTJs. Since the p-MTJ is a key device of current spintronics memory, i.e. spin-transfer-torque magnetoresistive random access memory (STT-MRAM), it attracts attention for further advance to explore new magnetic materials showing both large PMA and TMR. However, there have been no such materials other than FeCo(B)/MgO. Here, we report, for the first time, PMA in metastable bcc Co-based alloy, i.e. bcc CoMnFe thin films which are known to exhibit large TMR effect when used for electrodes of MTJs with the MgO barrier. The largest intrinsic PMAs were about 0.6 and 0.8 MJ/m3 in a few nanometer-thick CoMnFe alloy film and multilayer film, respectively. Our ab-initio calculation suggested that PMA originates from tetragonal strain and the value exceeds 1 MJ/m3 with optimizing strain and alloys composition. The simulation of the thermal stability factor indicates that the magnetic properties obtained satisfy the requirement of the data retention performance of X-1X nm STT-MRAM. The large PMA and high TMR effect in bcc CoMnFe/MgO, which were rarely observed in materials other than FeCo(B)/MgO, indicate that bcc CoMnFe/MgO is one of the potential candidates of the materials for X-1X nm STT-MRAM.

The Suppression Mechanisms of Passive Memory in Visual Working Memory: The Evidence from Electroencephalography.

Recent studies of visual working memory (VWM) underscore a structured hierarchy of storage states. Memories that are not immediately relevant to the task at hand but are essential for later use are transferred to a passive state, which operates independently of actively maintaining and manipulating current memories. Note that stimulating passive memory forcefully can reactivate it into an active state, resulting in a competition with active memory. Thus, to remain stable representations for both states within VWM, passive memory might involve sustained suppression during activity-silent maintenance to prevent reactivation from disrupting the current active storage. To investigate this, we analyzed lateralized EEG signals while human participants (both women and men) were engaged in a sequential presentation memory task across two experiments. The results revealed positive contralateral delayed activity components and lateralized alpha enhancement for passive memory, neural indicative of suppression on passive storage. In addition, the suppression effect was independent of the memory load in both the active and the passive states. These findings support the notion of sustained suppression during activity-silent maintenance of passive memory, facilitating the stable maintenance of distinct storage states and advancing our understanding of the dynamic coding framework in VWM.

Psychological Correlates of Post-COVID Condition

<span lang="SK">This article deals with psychological correlates of the post-COVID condition - depression, anxiety, and memory problems including remembering and forgetting. The work has a longitudinal character, while it examines the post-COVID condition in a group of participants based on infection of the disease Covid-19. At the same time, participants who never had this disease (N=86, for the 1. measure and N=38 for the 2. measure) are compared with those who were infected with COVID-19 (N=112, for the 1. measure and N=36 for the 2. measure), while participants were separated for analysis in groups based on the time from infection. Group of respondents who were infected within 3 months, from 3-9 months from infection and 9+ months since infecting with the disease. We used tests for comparison, such as the Kruskal Wallis test and multivariate procedures for within-subject and between-subject changes, using Anova for mixed experimental designs. We also determined the risk of being included in a group based on the time of infection with COVID-19 through multinominal logistic regression. We found differences in all variables in general between those infected with COVID-19 and those not. In neuropsychiatric aspects- anxiety and depression were confirmed cyclical features in those who were infected with COVID-19 in time, with the worst value of variables in the group from 3-9 months from infecting with the disease. In the current memory scale and its subscale, the act of remembering was a value that worsened over time and it was confirmed as an effect of group inclusion.</span>

Advanced spectroscopic methods for probing in-gap defect states in amorphous SiNx for charge trap memory applications

Silicon nitride (SiNx) serves as the charge trap layer in current 3D NAND flash memory devices. The precise formation mechanism and electronic structure of localized defect trap states in SiNx remain elusive. Here, we present a refined experimental methodology to elucidate the in-gap defect states and the band gaps in amorphous SiNx thin films. Our approach integrates high-resolution reflection electron energy loss spectroscopy (REELS) and spectroscopic ellipsometry (SE) for comprehensive analysis. By systematical analysis, we aim to provide a robust method for determining in-gap electronic states in SiNx. We investigated two different SiNx films prepared by plasma-enhanced chemical vapor deposition and sputtering. Our analysis revealed several distinct in-gap states and determined band gap energies. This approach not only provide advanced spectroscopic methods to characterize the defect electronic states in SiNx, but also applicable to other large band gap semiconductors or dielectrics to predict device-level characteristics for future devices.

SK 하이닉스의 HBM 경쟁력 요인과 향후 전략 분석

HBMs with higher performance and efficiency than current memory chips are in focus when AI comes. Due to the rapid increase in AI accelerator demand, semiconductor companies' interest in HBM is rising. Thanks to its yield, technology, material supply chain, and Nvidia partnership, SK Hynix remains number one in this industry. SK Hynix, which has failed to reverse the memory semiconductor business, is now writing a new history. Samsung Electronics, the second-largest company, is about a year behind SK. As always, competitors are trying to increase R&D spending, production capacity, and market share. This study examines SK Hynix's plan to maintain its No. 1 position in a constantly changing industry environment and how Samsung Electronics and Micron are attempting to overtake it. By thoroughly examining each competitor and the market environment, this paper provides short-term and mid- to long-term plans to help SK Hynix strengthen its market leadership.

Multicolor Fully Light-Modulated Artificial Synapse Based on P-MoSe2/PxOy Heterostructured Memristor.

Developing brain-inspired neuromorphic paradigms is imperative to breaking through the von Neumann bottleneck. The emulation of synaptic functionality has motivated the exploration of optoelectronic memristive devices as high-performance artificial synapses, yet the realization of such a modulatory terminal capable of full light-modulation, especially near-infrared stimuli, remains a challenge. Here, a fully light-modulated synaptic memristor is reported on a P-MoSe2/PxOy heterostructure formed by a facile one-step selenization process. The results demonstrate successful achievement of multiwavelength (visible 470 nm to near-infrared 808 nm) modulated switching operations (reset in 0.21-0.97 V) and diverse synaptic behaviors, including postsynaptic current, paired-pulse facilitation, short- and long-term memory (STM and LTM), and learning-forgetting. Notably, the device can exhibit a 3.42 μA PSC increase under six identical 655 nm stimuli, a 11.90-46.24 μA PSC modulation by changing 808 nm light intensity from 6 to 14 mW/cm2, and a transition from STM to LTM lasting between 2.47 and 4.27 s by a prolonged 808 nm pulse from 1 to 30 s. A novel possible light-induced switching mechanism in such a heterostructure is proposed. Furthermore, brain-like light-stimulated memory behavior and Pavlov's classical conditioning demonstrate the device's capacity for processing complex inputs. The study presents a design toward a multiwavelength modulated artificial visual system for color recognition.

Efficient clustering of large molecular libraries.

The widespread use of Machine Learning (ML) techniques in chemical applications has come with the pressing need to analyze extremely large molecular libraries. In particular, clustering remains one of the most common tools to dissect the chemical space. Unfortunately, most current approaches present unfavorable time and memory scaling, which makes them unsuitable to handle million- and billion-sized sets. Here, we propose to bypass these problems with a time- and memory-efficient clustering algorithm, BitBIRCH. This method uses a tree structure similar to the one found in the Balanced Iterative Reducing and Clustering using Hierarchies (BIRCH) algorithm to ensure time scaling. BitBIRCH leverages the instant similarity (iSIM) formalism to process binary fingerprints, allowing the use of Tanimoto similarity, and reducing memory requirements. Our tests show that BitBIRCH is already > 1,000 times faster than standard implementations of the Taylor-Butina clustering for libraries with 1,500,000 molecules. BitBIRCH increases efficiency without compromising the quality of the resulting clusters. We explore strategies to handle large sets, which we applied in the clustering of one billion molecules under 5 hours using a parallel/iterative BitBIRCH approximation.

α-CsPbI3 Quantum Dots ReRAM with High Air Stability Working by Valance Change Filamentary Mechanism.

The current memory system is facing obstacles to improvement, and ReRAM is considered a powerful alternative. All-inorganic α-CsPbI3 perovskite-based ReRAM working by electrochemical mechanism is reported, but the electrochemically active electrode raised difficulty in long-term stable operation, and bulk α-CsPbI3 device can not show resistive switching behavior with an inert metal top electrode. Herein, by making the α-CsPbI3 into QDs and applying it to the device with inert Au as the top electrode, the devices working by valence change mechanism are successfully fabricated. The large surface-to-volume ratio made an abundant amount of iodine vacancies and facile migration of vacancies allowed the device to work by valence change mechanism. The devices show reliable electrical characteristics, 800 cycles endurance and retention for over 4 × 104 s, and air stability for 1 month. This work demonstrates that applying the QDs can improve the stability and enable a new type of working mechanism in ReRAM.

Baťovany – Partizánske: The Contemporary Re/invention of the Heritage of a Baťa Company Town

The text focuses on the re/construction of identity and heritage conservation in the Slovak town of Partizánske. The town was founded as one of the industrial towns of the Baťa company at the turn of the 1930s and 1940s. Despite the efforts of experts, neither the town’s urban plan nor any significant part of it (except for the modernist church) is institutionally protected to this day. In this text, we offer an alternative approach to the re/construction of historical heritage and its institutional protection. In addition to qualities deemed valuable by art historians, this approach is informed by the current collective memory on which the town’s inhabitants base their relationship to historical heritage. We anticipate that focusing on the inhabitants’ current relationship and understanding of the city’s history and broadening the focus beyond the founding firm and its activities can support efforts to institutionalise local heritage conservation.

Ultra8T: A sub-threshold 8T SRAM with leakage detection

In energy-constrained scenarios such as IoT applications, the primary requirement for System-on-Chips (SoCs) is to increase battery life. However, when performing the sub/near-threshold operations, the relatively large leakage current hinders Static Random Access Memory (SRAM) from normal read/write functionalities at the lowest possible voltage (VDDMIN). In this work, we first propose a model that describes a specific relationship between read current and leakage noise in a given column. Based on the model, Ultra8T SRAM is designed to aggressively reduce VDDMIN by using a leakage detection strategy where the safety sensing time on bitlines is quantified without any additional hardware overhead. We validate the proposed Ultra8T using a 256 × 64 array in 28 nm CMOS technology. Post-simulation results show successful read operation at 0.25 V with 1.11 μs read delay, and the minimum energy required is 1.69 pJ at 0.4 V

Integration of history information Drives Serial Dependence and Stabilizes Working Memory Representations.

Serial dependence has shown seemingly contradictory effects on visual perception and working memory. While serial dependence promotes perpetual and mnemonic stability, it biases behavioral reports toward prior information. The neural mechanisms that drive both biasing and adaptive stabilizing effects are not well understood. We proposed and tested a reactivation and integration mechanism that can account for these contradictory effects. We used multivariate pattern analyses of EEG data (26 human participants, 17 females, 9 males) to examine the reactivation of prior reported orientation during the delay period of a visual working memory task. The reactivation strength of prior reports, but not prior sensory items, was predictive of the magnitude of serial dependency biases. These reactivated representations integrated with the representation of the current memory item and improved the ability to decode the current contents of memory. Overall, our data provide convergent evidence suggesting that prior reports in a visual working memory task are reactivated on the subsequent trial and become integrated with current memory representations. This similarity-dependent reactivation mechanism drives both report biasing and stabilization effects attributed to serial dependence in working memory.

Structural Transformer with Region Strip Attention for Video Object Segmentation

Memory-based methods in semi-supervised video object segmentation (VOS) achieve competitive performance by performing feature similarity between the current frame and memory frames. However, this operation involves two challenges: 1) instances of occlusion caused by interaction, and 2) interference from similar objects or clutters in the background. In this work, we propose a Structural Transformer with Region Strip Attention (STRSA) approach to address these challenges. Specifically, we build a Structural Transformer (ST) architecture to decompose the feature similarity between the long-term memory frames and the current frame into two aspects: a time–space part and an object significance part. This allows us to investigate the spatio-temporal relationship of pixels and capture the salient features of the objects. Therefore, the differences between pixels and the specificity of objects are fully exploited. In addition, we leverage the object location information from the long-term memory masks and the strip pooling to design a Region Strip Attention (RSA) module, which boosts the attention for the foreground regions and suppresses the background clutters. Extensive experiments on DAVIS, YouTube-VOS, and MOSE benchmarks prove that our method achieves satisfactory results and outperforms the retrained baseline model.

Age differences in the principal temporo-spatial components of EEG activity during a proactive interference task

Proactive interference (PI) is the disruptive effect of no longer relevant information on current working memory (WM) processing. PI effects in EEG data have been previously found to be altered in healthy aging, although it remains unclear the extent to which such changes reflect delayed or different brain mechanisms employed to overcome PI. Hence, we had twenty-six young (18–34 years) and sixteen old (53–68 years) healthy adults complete a Recent Probes task while EEG was recorded. Compared to young adults, old adults were slower, less accurate and less able to discriminate when they last saw a given stimulus, but PI effects on reaction time were greater in the former, likely due to a general difficulty that old adults had in the task. Temporo-spatial principal component analysis of the EEG data showed young and older adults to differ in terms of temporal and spatial characteristics of brain activity associated with resolving PI. YA showed a factor indicative of a medial frontal negativity (MFN) that showed greater amplitude in low compared to high PI trials. OA, in contrast, showed a late positive component (LPC), although similarly with larger amplitude in low compared to high PI trials. The modulation of the MFN component in YA may reflect the recruitment of cognitive control to overcome PI. The modulation of the LPC in OA may represent the detection of conflict between familiarity and context recollection during PI.

A quantum sensing metrology for magnetic memories

Magnetic random access memory (MRAM) is a leading emergent memory technology that is poised to replace current non-volatile memory technologies such as eFlash. However, controlling and improving distributions of device properties becomes a key enabler of new applications at this stage of technology development. Here, we introduce a non-contact metrology technique deploying scanning NV magnetometry (SNVM) to investigate MRAM performance at the individual bit level. We demonstrate magnetic reversal characterization in individual, <60 nm-sized bits, to extract key magnetic properties, thermal stability, and switching statistics, and thereby gauge bit-to-bit uniformity. We showcase the performance of our method by benchmarking two distinct bit etching processes immediately after pattern formation. In contrast to ensemble averaging methods such as perpendicular magneto-optical Kerr effect, we show that it is possible to identify out of distribution (tail-bits) bits that seem associated to the edges of the array, enabling failure analysis of tail bits. Our findings highlight the potential of nanoscale quantum sensing of MRAM devices for early-stage screening in the processing line, paving the way for future incorporation of this nanoscale characterization tool in the semiconductor industry.