Articles tagged with Memory Processes
A team of researchers has developed a new technique for magnetization switching in spintronic devices, nearly 100 times faster than current state-of-the-art methods. The breakthrough could lead to the development of ultrafast magnetic memory for computer chips that retain data even when power is off.
University of Pittsburgh professors Nathan Youngblood and Feng Xiong receive $380,000 to study phase-change materials for high-speed computing and optical storage applications. The project aims to overcome the challenges in electrically-controlled optical memory devices.
Researchers have discovered a memory effect that dramatically alters the Doppler wave signature in scattered waves. This effect, which appears in both relativistic and classical regimes, is influenced by memories of prior wave interactions, resulting in asymmetric peaks in the scattered spectrum.
Research at the University of Bristol discovers a new mechanism for learning in the brain that stabilises memories and reduces interference. This breakthrough provides insight into how humans form expectations and make accurate predictions about future events.
A Singapore-led collaboration created a hybrid simulator that maps complex phenomena of memristor memory technology with high accuracy. The simulator predicts transport behavior under various conditions, accounting for device characteristics and electron mobility, enabling more efficient designs.
A study by the University of Montreal has revealed the rules of synaptic plasticity, a process underlying learning and memory. The research found that dendritic spines, tiny protrusions on neurons, amplify or suppress incoming information based on its strength.
Researchers at Hokkaido University developed a hydrogel that mimics the human brain's dynamic memory function, encoding information that fades with time depending on intensity. The hydrogel's memory system can be programmed by temperature and learning time, allowing for stable memory establishment and controlled forgetting processes.
A team of Greek researchers evaluate the progress in optical memory domain over past 25 years, shedding light on physical mechanisms behind demonstrated devices and classifying performance metrics. They highlight benefits of different optical technologies and report recent achievements towards advancing memory functionalities.
The US Department of Energy has awarded Penn State over $10 million to develop new ferroelectric memory materials that can be stacked in the third dimension above processor chips. This breakthrough technology aims to mitigate the von Neumann bottleneck, allowing for seamless communication between memory and computation.
A nationwide project aims to engage 30,000 people in memory training to understand its cognitive benefits. The study will explore how different training methods affect individuals and shed light on the underlying mechanisms of brain training.
Researchers at Bar-Ilan University have developed a structure that can support exponentially many discrete magnetic states, opening the door to ultra-high-density magnetic memory and novel computing applications.
A recent study reveals a new magnetic state called skyrmion, which can be manipulated using electric fields to create more compact and efficient nanomagnetic devices. This breakthrough could lead to significant improvements in computer performance and reduce energy consumption.
The research team developed a new MTJ stack design technology and highly reliable fabrication technology for Quad interface type iPMA-MTJ. Quad-MTJ achieved a 20% reduction in write density at a 10ns high speed write operation, maintaining data retention for over 10 years and higher operating temperatures than Double-MTJ.
Scientists develop novel 3D-spiral circuit design for faster and more energy-efficient machine learning training. The design reduces interconnecting wiring needs and enhances algorithm accuracy up to 90%. This innovation may enable the next generation of energy-efficient AI devices.
Rice University researchers have demonstrated methods to optimize data-centric processing, improving energy efficiency by up to two orders of magnitude. Two complementary approaches, TIMELY and SmartExchange, were presented at ISCA 2020, leveraging innovative hardware architecture and co-designing with machine-learning algorithms.
Researchers at the Picower Institute for Learning and Memory discovered that the lateral septum tracks rewards with an emphasis on location. This brain region plays a crucial role in filtering and converting raw information about location into reward-specific output.
Sandia will be one of the first DOE laboratories to receive Fujitsu's new A64FX processor, optimized for memory-speed bottleneck breakage. The 48-core processor provides greater fractions of usable peak performance and supports collaboration with the Japanese supercomputing community.
Researchers found that immune receptor TLR4 suppresses neuronal activity in the normal brain to improve signal-to-noise ratio, while increasing excitability after injury. Inhibiting TLR4 signaling reduces long-term working memory deficits and improves memory function weeks to a month after brain injury.
The ability to extinguish fearful memories relies on DNA flexibility, which enables rapid changes in gene activity and RNA editing. Researchers found that the enzyme ADAR1, which converts Z-DNA back into B-DNA, plays a critical role in fear extinction.
A study found that age-related hearing loss damages the brain's hippocampus, a key memory organ. The findings suggest that progressive hearing impairment impairs memory formation by altering neurotransmitter receptors.
A new study found that people are no better than chance at identifying false memories of committing a crime, even when told they're fake. The research used videos from a previous study where participants were tricked into believing they had committed a crime, and found that many true memories looked like false ones.
Researchers identified genes ABTB1 and GRB10 as influential in nutrient-sensing pathways, associated with performance on memory tasks. Lifestyle changes, such as diet and exercise, can delay memory decline, but genetic variations affect their effectiveness.
A recent study identified distinct gene signatures associated with memory creation, found in cortical and subcortical regions of the brain. These genes play crucial roles in memory processes, immune signaling, neuronal generation, and mRNA production.
Researchers at Tohoku University have developed a high-speed spin-orbit-torque magnetoresistive random access memory cell compatible with 300mm Si CMOS technology. The SOT-MRAM cell achieves high-speed switching and robust thermal stability, making it suitable for replacing SRAM in commercial applications.
Researchers created a compact device to interface with electronic architecture of traditional computers using light-encoded information. The device uses surface plasmon polariton to confine light into nanoscopic dimensions, overcoming the incompatibility between electrical and optical computing.
Scientists have developed an atomic-scale manufacturing tool that speeds up data storage and increases computing power for artificial intelligence. This breakthrough could eliminate a gigatonne of carbon emissions while improving data capacities.
A team of researchers from Tokyo Tech has developed a novel memory device that can switch between three states using nickel-based layers, leading to extremely low energy consumption and faster performance. The device is inspired by solid lithium-ion batteries and boasts a miniature battery-like architecture.
Researchers developed lipid-based memcapacitors that mimic biological synapses, accelerating routes to neuromorphic computing. The discovery could support the emergence of biology-inspired computing networks for sensory approaches to machine learning.
Researchers at RUDN University have developed a method to identify electronic devices using defects in flash memory cells, which are unique to each device. The distribution and nature of these defects can be used to create 'fingerprints' that cannot be faked.
A team led by Keren Bergman is developing on-chip integrated photonic devices for interconnects that consume 100 times less energy than today's communication systems while delivering 100 times more communication bandwidths between computing nodes. The project aims to break the boundaries on available communications capabilities of comp...
A research group led by Emery Berger has developed Mesh, a system that can automatically reduce memory demands in applications. By taking advantage of virtual memory hardware, Mesh effectively squeezes out memory gaps, resulting in significant reductions of up to 40% in memory usage for certain applications.
Researchers developed a novel liquid process to fabricate an affordable multiferroic nanocomposite film, exhibiting a strong correlation between its electric and magnetic properties. This breakthrough enables the production of materials for various applications such as large-volume memory, spatial light modulators, and unique sensors.
A team of researchers at HSE University found that working memory stores objects and their features independently, making it easier to remember single-object information. This hierarchical structure suggests that working memory can store multiple types of information simultaneously.
The National Science Foundation is funding a new supercomputer at the Pittsburgh Supercomputing Center (PSC) with a $10-million grant. The Bridges-2 system will provide massive computational capacity and flexibility to researchers in Pennsylvania and nationwide.
A study from MPFI has uncovered that CaMKII decodes calcium signals predominantly through its autonomous activity, rather than just its interaction with calmodulin. This finding broadens our understanding of how molecules contribute to memory and synaptic plasticity.
University of Houston researcher Margaret Cheung is exploring the molecular structure of memories at the precise moment they are formed in single neurons. She aims to understand how calcium signaling and calmodulin affect cellular processes, including human cognition.
Research reveals two distinct brain regions involved in recognizing raw and processed foods, highlighting the importance of semantic memory in identifying key characteristics. The study's findings have significant clinical implications, particularly for understanding eating disorders related to neurodegenerative diseases.
The team's artificial synapse is similar to a battery, emulating how learning is wired in the brain, and processes data in one action. The prototype array outperformed expectations with high speed, energy efficiency, reproducibility and durability, paving the way for small devices to support artificially intelligent learning.
A study found that a wider network of brain regions than previously thought is engaged in remembering a familiar face, linking the 'face network' with social, visual, and auditory circuits. This integration predicts facial memory performance.
Researchers found that piriform cortex stimulation triggers synaptic plasticity in the hippocampus, leading to enhanced memory formation. The study provides a theoretical basis for understanding the special role of olfaction in memory formation and retrieval.
Researchers at U of T discovered the human hippocampus is sensitive to time information on short timescales, helping bind together information about time when forming long-term memories. The study builds on a previous rodent study that found 'time cells' in the hippocampus.
Research reveals distinct brain regions process concrete and abstract word meanings separately, with a network of areas working together in different languages. The findings provide new insights into how our brains represent word-meaning structure.
The study used EEG and ERP to assess neural activity during simultaneous interpretation of continuous prose. The results showed that interpreters work in a dynamic mode of redistribution of attention, with increased cognitive resources engaged for working memory when lagging behind the speaker.
Researchers found a new network of brain regions involved in temporal memory, which stores information about when events happen. This discovery may further understanding of age-related dementia and Alzheimer's disease.
Researchers at Tohoku University have developed a 128Mb-density STT-MRAM with a write speed of 14 ns, setting a new world record. This technology has the potential to significantly reduce power consumption in embedded memory applications.
Research at Ruhr-University Bochum shows that blindness disrupts brain's organisation and memory ability in mice. After vision loss, other senses like touch, hearing, and smell become more sensitive over time.
A team of researchers developed a neuroinspired hardware-software co-design approach that can make neural network training more energy-efficient and faster. The approach uses a type of energy-efficient neural network called spiking neural networks, combined with the soft-pruning algorithm to minimize computing power and time.
A WSU research team has uncovered previously unknown vulnerabilities in high-performance computer chips, which could be exploited by malicious attacks. The communications system is a critical component of the chip, and its malfunction can have severe consequences.
Researchers found that music and pictures evoke strong positive emotions due to personal memories, especially for social and positive feelings like tenderness and joy. The study also showed that personal memories play a crucial role in triggering sad emotions in response to music and pictures, particularly for dementia patients.
Researchers at Kyoto University successfully created intense terahertz pulses to fine-tune the switching behavior of a phase-change memory material. This breakthrough could lead to faster and more stable memory technologies with increased density.
Researchers have developed memristors that mimic the brain's ability to process information efficiently and adapt to changing conditions. The new devices, made from niobium-doped strontium titanate, can 'forget' and retain memories, making them suitable for future electronic brains.
The MIT CSAIL team has developed a new approach called DAWG to securely partition cache memory, making it harder for hackers to exploit Meltdown and Spectre vulnerabilities. This method can have immediate applications in cloud computing, especially in fields like medicine and finance.
Researchers at IBM developed a new computer architecture with co-located memory and processing, significantly outperforming conventional computers. This brain-inspired design achieved 200 times faster performance in machine learning tasks.
Research found that short periods of light exercise boost functional connectivity in the hippocampus, a region critical for memory processing. Participants showed improved memory performance after exercise, with increased connectivity correlating to better results.
Researchers at IMDEA Networks Institute are exploring a novel 'DisCoEdge' system to optimize device usage and create new services. The system aims to spread heavy computational tasks and large storage over simple devices, enabling novel applications like cooperative data sharing and corporate storage systems.
A team of researchers from the Moscow Institute of Physics and Technology has developed a model for predicting hand movement trajectories based on electrocorticogram data. This technology could enable exoskeletons that allow patients with impaired mobility to regain movement.
Song Jiang is working on a three-year grant to improve cache efficiency in software, allowing for faster data access and large memory usage. He aims to develop prototypes that can combine or skip steps to access data automatically without slowing CPU speeds.
Researchers have developed a memory-processing unit using memristors, which can perform numerical calculations in parallel, reducing the need for manual multiplication and summation. This innovation has potential applications in machine learning, artificial intelligence, and simulations for weather forecasting.
Scientists aim to create a DNA-based read-only memory (ROM) with applications in electronic devices, leveraging the unique electrical properties of DNA molecules. The proposed technology could lead to more efficient and compact devices.
The FAU team's award-winning paper focuses on improving the accuracy of ECM performance model and multicore power model to better describe energy properties of processors commonly used in HPC resources. The improved models provide a more accurate view of processors' actual performance under high computational demands.