Figure - available from: Advanced Materials
This content is subject to copyright. Terms and conditions apply.
a) Schematic illustration of forgetting curve, proposed by Ebbinghaus. b) Schematic illustration of memory loss in biological brain.
Source publication
The nature of repetitive learning and oblivion of memory enables humans to effectively manage vast amounts of memory by prioritizing information for long-term storage. Inspired by the memorization process of the human brain, an artificial synaptic array is presented, which mimics the biological memorization process by replicating Ebbinghaus' forget...
Similar publications
Biological visual systems have inspired the development of various artificial visual systems including those based on human eyes (terrestrial environment), insect eyes (terrestrial environment) and fish eyes (aquatic environment). However, attempts to develop systems for both terrestrial and aquatic environments remain limited, and bioinspired elec...
Citations
... These artificial synapses serve as hardware platforms for neural networks that can learn, remember, and generalize from data through deep learning algorithms [9][10][11][12]. The study of synaptic devices, driven by multidisciplinary efforts in materials science, electronics, and neuroscience, has resulted in a diverse array of technological approaches [13][14][15][16][17][18]. The study of synaptic devices, driven by multidisciplinary efforts in materials science, electronics, and neuroscience, has resulted in a diverse array of technological approaches [13][14][15][16][17][18][19][20][21]. ...
... The study of synaptic devices, driven by multidisciplinary efforts in materials science, electronics, and neuroscience, has resulted in a diverse array of technological approaches [13][14][15][16][17][18]. The study of synaptic devices, driven by multidisciplinary efforts in materials science, electronics, and neuroscience, has resulted in a diverse array of technological approaches [13][14][15][16][17][18][19][20][21]. ...
... Indeed, researchers in South Korea have conducted active research on applying neuromorphic computing for artificial synapses, 45 resistive switching and memristor devices, 46,47 and synaptic plasticity and learning algorithms. 48 In the UK, there have been active discussions on decarbonization strategies, 15 renewable energy, 17,49 and energy transition and environmental impacts. 16,50 For the US, a high portion of its main topics are related to structural materials such as "cement paste" and "composite modeling". ...
Materials science research has garnered extensive attention from industry, society, policy, and academia. However, understanding the research landscape and extracting strategic insights are challenging due to the increasing diversity and volume of publications. This study proposes a natural language processing-based protocol for extracting text-encoded topics from a large volume of scientific literature, uncovering research interests of scientific communities, as well as convergence trends. We report a topic map, representing the materials science research landscape with text-mined 257 topics regarding biocompatible materials, structural materials, electrochemistry, or photonics. We analyze the topic map in terms of national research interests in materials science, revealing competitive positions and strategies of active nations. For example, it is found that the increasing trend of research interest in machine learning topic was captured in the United States earlier than other nations. Similarly, our journal-level analyses serve as reference information for journal recommendations and trend guidance, showing that the main topics and research interests of materials science journals slightly changed over time. Moreover, we build the topic association network which can highlight the status and future potential of interdisciplinary research, revealing research fields with high centrality in the network such as machine learning-enabled composite modeling, energy policy, or wearable electronics. This study offers insightful results on current and near-future materials science research landscapes, facilitating the understanding of stakeholders, amidst the fast-evolving and diverse knowledge of materials science.
... The relationship between memory retention and time corresponds to the Ebbinghaus forgetting curves 45,46 , and repeating the learning process accompanied by forgetting experiences (Fig. 4f) can convert short-term memory (STM) to long-term memory (LTM). In this experiment (Fig. 4g), the gray dash line represents the learning target level, and the PAS is exposed to two types of 395 nm CPL for studying polarization-dependent learnability. ...
Circularly polarized light (CPL) adds a unique dimension to optical information processing and communication. Integrating CPL sensitivity with light learning and memory in a photonic artificial synapse (PAS) device holds significant value for advanced neuromorphic vision systems. However, the development of such systems has been impeded by the scarcity of suitable CPL active optoelectronic materials. In this work, we employ a helical chiral perovskite hybrid combined with single-wall carbon nanotubes to achieve circularly polarized ultraviolet neuromorphic vision sensing and imaging. The heterostructure demonstrates long-term charge storage as evidenced by multiple-pulsed transient absorption measurements and highly sensitive circular polarization-dependent photodetection, thereby enabling efficient CPL-resolved synaptic and neuromorphic behaviors. Significantly, our PAS sensor arrays adeptly visualize, discriminate, and memorize distinct circularly polarized images with up to 93% recognition accuracy in spiking neural network simulations. These findings underscore the pivotal role of chiral perovskites in advancing PAS technology and circular polarization-enhanced ultraviolet neuromorphic vision systems.
... Voltage pulses of 100 ms with an amplitude of 7 V were applied several times to a Cu-based bilayer device, and the retention time of the device was measured by reading the device conductance at a pulse with an amplitude of 1 V at equal time intervals. The memory decay behaviors resembled Ebbinghaus's forgetting curve, where short-term memory (STM) decay in an early stage of the retention test was dominant, followed by long-term memory (LTM) decay in a later stage 50 . As shown in Fig. 7b, the maximum conductance value increased as the number of pulses applied increased. ...
Various memristive devices have been proposed for use in neuromorphic computing systems as artificial synapses. Analog synaptic devices with linear conductance updates during training are efficiently essential to train neural networks. Although many different analog memristors have been proposed, a more reliable approach to implement analog synaptic devices is needed. In this study, we propose the memristor of a Cu/SiO x /implanted a-SiGe x /p ⁺⁺ c-Si structure containing an a-Si layer with properly controlled conductance through Ge implantation. The a-SiGe x layer plays a multifunctional role in device operation by limiting the current overshoot, confining the heat generated during operation and preventing the silicide formation reaction between the active metal (Cu) and the Si bottom electrode. Thus, the a-SiGe x interface layer enables the formation of multi-weak filaments and induces analog switching behaviors. The TEM observation shows that the insertion of the a-SiGe x layer between SiO x and c-Si remarkably suppresses the formation of copper silicide, and reliable set/reset operations are secured. The origin of the analog switching behaviors is discussed by analyzing current-voltage characteristics and electron microscopy images. Finally, the memristive-neural network simulations show that our developed memristive devices provide high learning accuracy and are promising in future neuromorphic computing hardware.
... The most important factor of the formula is time. The original experiments of Ebbinghaus were since then replicated, confirming the correctness of the formula with some small modification in its smoothness [13], and it has also been investigated in the context of brain function [16]. The psychological experiments of Averell and Heathcote [2] confirmed that the exponential curve is the best fit to model human forgetting. ...
"Program comprehension is a continuously important topic in computer science since the spread of personal computers, and several program comprehension models have been identified as possible directions of active code comprehension. There has been little research on how much programmers remember the code they have once written. We conducted two experiments with a group of Computer Science MSc students. In the first experiment, we examined the code comprehension strategies of the participants. The students were given a task to implement a minor feature in a relatively small C++ project. In the second experiment, we asked the students 2 months later to complete the same task again. Before starting the clock, we asked the students to fill a questionnaire which aimed to measure program code-related memory retention: we inquired about how much the students remembered the code, down to the smallest relevant details, e.g. the name of functions and variables they had to find to complete the task. After the second experiment, we could compare the solution times of those students who participated in both parts. As one result, we could see that these students could solve the task in shorter time than they did in the first experiment. We also looked at the results of the questionnaire: the vast majority of students could not precisely remember more than two or three identifiers from the original code. In this paper, we will show how this result compares to the forgetting curve. 2010 Mathematics Subject Classification. 68U99. 1998 CR Categories and Descriptors. I.m [Computing Methodologies]: Miscellaneous; J.m [Computer Applications]: Miscellaneous. Key words and phrases. code comprehension, memory retention, experiment."
... The human eye and visual cortex mimetic chips are the most powerful devices for AI-embedded hardware-based neuromorphic computing [144][145][146][147]. Real-time weight updates from input signals provided by light-based patterned images, as performed by the human eye, and signal processing in a neural network (human brain) need to be developed for practical applications [148][149][150][151]. Although AI chips composed of light sensors and neural systems still face challenges emulating biological neurons in the human eye and brain, several state-of-the-art methods have been developed for the integration of desired components [152][153][154][155][156]. ...
Memristors mimic synaptic functions in advanced electronics and image sensors, thereby enabling brain-inspired neuromorphic computing to overcome the limitations of the von Neumann architecture. As computing operations based on von Neumann hardware rely on continuous memory transport between processing units and memory, fundamental limitations arise in terms of power consumption and integration density. In biological synapses, chemical stimulation induces information transfer from the pre- to the post-neuron. The memristor operates as resistive random-access memory (RRAM) and is incorporated into the hardware for neuromorphic computing. Hardware composed of synaptic memristor arrays is expected to lead to further breakthroughs owing to their biomimetic in-memory processing capabilities, low power consumption, and amenability to integration; these aspects satisfy the upcoming demands of artificial intelligence for higher computational loads. Among the tremendous efforts toward achieving human-brain-like electronics, layered 2D materials have demonstrated significant potential owing to their outstanding electronic and physical properties, facile integration with other materials, and low-power computing. This review discusses the memristive characteristics of various 2D materials (heterostructures, defect-engineered materials, and alloy materials) used in neuromorphic computing for image segregation or pattern recognition. Neuromorphic computing, the most powerful artificial networks for complicated image processing and recognition, represent a breakthrough in artificial intelligence owing to their enhanced performance and lower power consumption compared with von Neumann architectures. A hardware-implemented CNN with weight control based on synaptic memristor arrays is expected to be a promising candidate for future electronics in society, offering a solution based on non-von Neumann hardware. This emerging paradigm changes the computing algorithm using entirely hardware-connected edge computing and deep neural networks.
... Voltage pulses of 100 ms with an amplitude of 7V was applied several times to a Cu-based bi-layer device, and the retention time of the device was measured by reading the device conductances at a pulse with an amplitude of 1V at equal time intervals. The memory decay behaviors resemble Ebbinghaus's forgetting curve, where short term memory (STM) decay in an early stage of the retention test is dominant, followed by long term memory (LTM) decay in a later stage 50 . As shown in Fig. 7(b), the maximum conductance value increases as the number of pulses applied increases. ...
Various memristive devices have been proposed for use in neuromorphic computing systems as artificial synapses. The analog synaptic devices with linear conductance updates during training are essential to train neural networks efficiently. Although many different analog memristors have been proposed, a more reliable approach to implement the analog synaptic devices are required. In this study, we propose the memristor of a Cu/SiO x /implanted a-SiGe x /p ⁺⁺ c-Si structure containing a-Si layer with properly controlled conductance through Ge implantation. The a-SiGe x layer plays a multi-functional role in the device operation by limiting current overshoot, confining heat generated during operation and preventing silicide formation reaction between active metal (Cu) and the Si bottom electrode. Thus, the a-SiGe x interface layer enables the formation of multi-weak filaments and in turn induce analog switching behaviors. The TEM observation reveals the insertion of the a-SiGe x layer between SiO x and c-Si suppresses remarkably the formation of copper silicide, and the reliable set/reset operations were secured. The origin of the analog switching behaviors was discussed by analyzing current-voltage characteristics and electron microscopy images. Lastly, the memristive-neural network simulations showed that the memristive devices developed in this study provide a high learning accuracy and be promising in future neuromorphic computing hardware.
... Although efforts have been made to develop efficient control systems, current CMOS-based robotic control systems have inherent limitationsthe von Neumann bottleneck and the binary logic structureresulting in signal delay and poor chip integration [6][7][8][9][10][11][12] . The advent of synaptic transistors provided a breakthrough in addressing signal delay and chip integration issues [13][14][15][16][17] ; these devices are capable of parallel computation and analog signal processing, similar to the human nervous system [18][19][20][21][22] . Among the various types of synaptic transistors 6,12,[23][24][25][26][27] , ion-based electrochemical synaptic transistors are attracting attention for biomimetic applications since their working mechanism is similar to that of human synapses. ...
... Among the various types of synaptic transistors 6,12,[23][24][25][26][27] , ion-based electrochemical synaptic transistors are attracting attention for biomimetic applications since their working mechanism is similar to that of human synapses. Human synapses transmit biological signals by releasing neurotransmitters from presynaptic neurons; the neurotransmitters pass through the synaptic cleft and reach postsynaptic neurons 18,[20][21][22][28][29][30] . In the case of an electrochemical synaptic transistor, signal transmission occurs through the electrical-input-signal-induced penetration of a semiconducting channel by the ionic species. ...
With advances in robotic technology, the complexity of control of robot has been increasing owing to fundamental signal bottlenecks and limited expressible logic state of the von Neumann architecture. Here, we demonstrate coordinated movement by a fully parallel-processable synaptic array with reduced control complexity. The synaptic array was fabricated by connecting eight ion-gel-based synaptic transistors to an ion gel dielectric. Parallel signal processing and multi-actuation control could be achieved by modulating the ionic movement. Through the integration of the synaptic array and a robotic hand, coordinated movement of the fingers was achieved with reduced control complexity by exploiting the advantages of parallel multiplexing and analog logic. The proposed synaptic control system provides considerable scope for the advancement of robotic control systems.
... For each memory path in S(e P m ,t e pair ), we utilize the Ebbinghaus curve to model its personal dynamics. Inspired by [13] and [17], we consider three factors to model personal memory path intensity: unconscious memory intensity, conscious memory intensity, and time decay. The unconscious memory intensity is the user's general impression of the memory path. ...
Nowadays, many dynamic recommendations still suffer from the insufficiency of finding user online interest evolving patterns because of those complicated interactions. In general, each interaction is usually impacted by multiple underlying reasons, which needs us to open the “box” of each interaction instance instead of simply treating them as a pair-wise link. Besides, different users usually perform differently for their long-term and short-term tastes, leaving traditional sequential models far from personalized. In this article, we propose a novel recommendation model based on Graph Diffusion and Ebbinghaus Curve. Specifically, to explore the underline reasons for different interactions, we explore an underlying sub-graph for each interaction and find important reasoning paths within the sub-graph via a well-designed graph diffusion method. To capture users’ personalized strategies on long-term and short-term tastes, we are inspired by the Ebbinghaus Curve, which can naturally describe users’ memory patterns, and design an effective neural network to process users’ evolving behaviors. We conduct extensive experiments on four real-world datasets and the results further confirm the superiority of our model compared with existing state-of-the-art baselines.
... Correspondingly, if a neuron within the sensory field in visual cortex receives two competing stimuli, the attended stimulus has an advantage over the unattended stimulus 5 . Inspired by biological system, previous efforts on hardware implementation of selective attention 1,6,[10][11][12][13][14] , are based on CMOS and conventional transistors, whereas it takes up large footprint and high computational cost. Meanwhile, the sensory unit is separated from processing system, which leads to tremendous challenge to synchronously handle signals. ...
Selective attention is an efficient processing strategy to allocate computational resources for pivotal optical information. However, the hardware implementation of selective visual attention in conventional intelligent system is usually bulky and complex along with high computational cost. Here, programmable ferroelectric bionic vision hardware to emulate the selective attention is proposed. The tunneling effect of photogenerated carriers are controlled by dynamic variation of energy barrier, enabling the modulation of memory strength from 9.1% to 47.1% without peripheral storage unit. The molecular polarization of ferroelectric P(VDF-TrFE) layer enables a single device not only multiple nonvolatile states but also the implementation of selective attention. With these ferroelectric devices are arrayed together, UV light information can be selectively recorded and suppressed the with high current decibel level. Furthermore, the device with positive polarization exhibits high wavelength dependence in the image attention processing, and the fabricated ferroelectric sensory network exhibits high accuracy of 95.7% in the pattern classification for multi-wavelength images. This study can enrich the neuromorphic functions of bioinspired sensing devices and pave the way for profound implications of future bioinspired optoelectronics. Selective attention is an efficient processing strategy to allocate computational resources for pivotal optical information. Here, the authors propose a bionic vision hardware to emulate the behavior, showing a potential in image classification.
