Articles tagged with Computer Architecture
Researchers developed a new type of nanoelectronic device mimicking the human brain's efficient neuron connections, reducing energy consumption for AI systems. The hafnium-based devices achieve switching currents millions of times lower than conventional devices and store programmed states for around a day.
The ARU Arm AI Lab will provide researchers and students with access to advanced Arm AI technology, focusing on real-world applications in healthcare and life sciences. This partnership will also support emerging talent and drive innovation, building on existing collaborations and industry projects.
Researchers have developed a robotic gripper called ChicGrasp, which can learn to handle chickens by imitating human movements. The system uses advanced imitation learning algorithm and camera perceptions to grasp a chicken carcass by the legs and lift it on a shackle conveyor.
Study by Duke engineers reveals that a common device architecture used to test 2D transistors overstates their performance prospects up to sixfold. The back-gated architecture amplifies the transistor's performance using contact gating, but has physical limitations that prevent its use in commercial technologies.
The institute will focus on technologies required to operate next-generation AI supercomputer systems reliably while accelerating scientific breakthroughs. It aims to drive innovation in making AI data centers more efficient, reliable, secure, and integrated with the nation's energy system.
Researchers at Adelaide University have developed a world-first cybersecurity system to protect drones from hacking, signal disruption, and malicious software. The system uses Software-Defined Wide Area Networking (SD-WAN) to act as a smart traffic controller for internet connections, making drones less vulnerable to cyber threats.
Researchers will gather to discuss how to engineer AI systems that work in the real world, focusing on composition, optimization, verification, and evaluation. The conference aims to establish shared foundations for a new class of software, including methods for evaluating models and ensuring durability, efficiency, and dependability.
New research demonstrates neuromorphic computing's ability to solve complex mathematical problems, including partial differential equations. This technology has the potential to revolutionize energy-efficient computing and tackle real-world challenges.
Researchers developed a bio-inspired neuron platform that processes and learns information using light and electronics integrated on a single platform. The chip achieves 92% image recognition accuracy and demonstrates key synaptic behaviors found in biological learning.
The UC3M's new supercomputer is a state-of-the-art facility designed to support research and development projects requiring high-performance computing. The system features over 11,500 CPU cores and 42 GPUs, enabling the processing of large volumes of data and complex numerical simulations.
The new Spectra supercomputer at Sandia National Laboratories features a unique chip architecture that prioritizes tasks in real time, promising increased performance and reduced power consumption. Researchers will push the limits of this technology to simulate complex national security tasks.
A team of researchers developed a groundbreaking Full Earth System Simulation at 1 km resolution, capturing energy, water, and carbon flow through the atmosphere, ocean, and land. This innovation has enormous potential to provide detailed global information on local scales about future warming implications.
A new Australian study computes the creative ability of LLMs using standard mathematical principles, showing a maximum of 0.25 on a scale from 0 to 1. The findings challenge widespread assumptions about AI's creative powers and offer clarity amid a global debate.
A small-scale neuromorphic computer prototype learned patterns and made predictions using fewer training computations than conventional AI systems. The prototype, developed by the University of Texas at Dallas team, integrates memory storage with processing, allowing for more efficient AI operations with lower costs.
Scientists from NTU Singapore propose building carbon-neutral data centres in Low Earth Orbit, harnessing unlimited solar energy and natural radiative cooling. This concept offers a sustainable environment for computing with global scalability and minimal land constraints.
The Stowers Institute has appointed its first AI Fellow, Sumner Magruder, to harness the potential of artificial intelligence in biological research. He will collaborate with researchers to design new algorithms and unlock insights from large datasets.
VFF-Net applies label-wise noise labelling, cosine similarity-based contrastive loss, and layer grouping to improve image classification performance compared to conventional forward-forward networks. The algorithm reduces test errors on various datasets, enabling lighter and more brain-like training methods that make AI more sustainable.
Amarasinghe recognized for groundbreaking work in domain-specific languages and exceptional mentorship, advancing the global computing community. His contributions have successfully bridged the gap between software and hardware to fully exploit modern hardware resources.
Researchers at the University of Missouri have developed an AI-powered method to detect hidden hardware trojans in chip designs, offering a 97% accurate solution. The approach leverages large language models to scan for suspicious code and provides explanations for detected threats.
A new international project aims to protect fragile quantum information from decoherence and loss, a key barrier to quantum computing's progression. The Magenium qubit design stores information in small, symmetric clusters of qubits, potentially allowing quantum data to last significantly longer than current methods.
Ana Veroneze Solórzano recognized for broadening HPC's societal impact with privacy-preserving mechanisms, while Yafan Huang advances exascale computing with ultra-fast compression algorithms. Both receive fellowships to support their research on high-performance computing applications.
A $5 million NSF grant is powering AI innovations in Pegasus, a widely used workflow management network. Researchers aim to develop intelligent workflow management tools that adapt to evolving user needs and dynamic resource conditions.
This book provides a beginner-friendly resource on the impact and evolution of decentralized networks, highlighting their applications in healthcare, supply chains, agriculture, climate monitoring, and education. The authors emphasize sustainability, data security, and ethical tech adoption.
Researchers developed an open-source software tool, Phoenix, to simulate light behavior in quantum systems, solving wave equations in record time without high-performance computing expertise. The program is up to a thousand times faster and 99.8% more energy-efficient than conventional tools.
The team aims to deliver AI power directly to devices, improving resilience and speed in constrained environments. By processing data step-by-step across a network of devices, they can create a safe and adaptable system that can withstand attacks and extreme conditions.
The new compute roadmap aims to improve lives and livelihoods with advanced computing power. It includes investments in AI research, healthcare diagnostics, and renewable energy.
Skia identifies and decodes shadow branches, storing them in a memory area to alleviate bottlenecks and improve throughput. The technique can lead to quicker performance and less power consumption for data centers.
César A. Uribe, a Rice University professor, has received an NSF CAREER Award to develop mathematical tools for decentralized learning in AI and data science. His research aims to create more efficient networks of computers that can process massive amounts of data without relying on centralized coordination.
Purdue researchers create a new method to notify CPUs without polling, reducing interrupt overhead and improving efficiency. This discovery impacts cloud systems and large data centers, driving smooth operations throughout networks.
A new approach to AI developed by Texas A&M University engineers mimics the human brain's neural processes, integrating learning and memory in a single system. This 'Super-Turing AI' has the potential to revolutionize the industry by reducing energy consumption and environmental impact.
Researchers at MIT created a photon-shuttling interconnect that facilitates remote entanglement, a key step toward developing practical quantum computers. The device enables all-to-all communication between multiple superconducting quantum processors, paving the way for more efficient and scalable quantum computing.
A study developed by the US Department of Energy's Thomas Jefferson National Accelerator Facility aims to lower data center costs using machine learning. The Digital Data Center Twin (DIDACT) system detects anomalies and diagnoses their source using AI continual learning, reducing downtime for scientists processing data from experiments.
Researchers propose several key features to optimize sparsity, massive parallelism, and hierarchical structure in neural representation for neuromorphic systems. The goal is to achieve energy efficiency and compactness while retaining information at high fidelity.
Researchers developed a groundbreaking photonic platform to overcome limitations in in-memory computing, enabling faster calculations and greater efficiency. The innovative magneto-optical memories consume about one-tenth the power of traditional electronics and can be rewritten billions of times.
Mayo Clinic's Digital Pathology platform is transforming pathology by leveraging large datasets and AI models for faster, more accurate diagnoses. The partnership with NVIDIA and Aignostics aims to improve patient care and accelerate medical breakthroughs.
The Department of Energy's new research centers, led by SLAC National Accelerator Laboratory, aim to make microelectronics more energy efficient and operate in extreme environments. Researchers will focus on innovating material design, devices, and systems architectures to push computing and sensing capabilities.
A team of researchers has developed a novel technique to steal artificial intelligence (AI) models by monitoring electromagnetic signals. The method allows attackers to recreate the high-level features of an AI model with 99.91% accuracy, potentially undermining intellectual property rights and exposing sensitive data.
A new study by Lancaster University reveals that language in job ads can unintentionally reinforce or disrupt labour force gender/racial composition. Workforces with more women tend to use family-friendly policies in ads, while racial minority workers' ads lack impact, the research shows.
Researchers at Tokyo University of Science have developed a new method called black-box forgetting, which enables selective removal of unnecessary information from large pre-trained AI models. This approach enhances model efficiency and improves privacy by reducing computational resources and information leakage.
A three-dimensional quantum error correction architecture was discovered, which can handle errors scaling like L<sup>2</sup> (LxL) in two-dimensions. This breakthrough promises to enhance the reliability of quantum information storage and reduce physical computing resources needed for 'logical qubits', paving the way for a more compact
Researchers from Pitt, UC Santa Barbara, University of Cagliari, and Institute of Science Tokyo have developed a new method for photonic in-memory computing that combines non-volatility, multibit storage, high switching speed, low switching energy, and high endurance in a single platform.
Researchers at the University of Tokyo introduce a new optical computing scheme called diffraction casting, which improves upon existing methods. The system uses light waves to perform logic operations and has shown promise in running complex calculations, including those used in machine learning.
The U.S. Department of Energy has selected Kate Petersen Mace as the project director for the High Performance Data Facility (HPDF), a first-of-its-kind project providing resources for data-intensive science. Mace will lead the team in developing transformational capabilities to serve the DOE community.
Researchers developed a real-time stream-based data compression technology that automatically detects frequently occurring data patterns and compresses them to minimize data volume. This innovation enables high-speed, compact data compression modules in hardware without additional processors or memory.
Researchers at North Carolina State University developed a mechanical computer that uses a complex structure of rigid polymer cubes to store, retrieve and erase data. The system has reversible features allowing users to control when data editing is permitted.
Two new Parnas Fellows, Prof. Hongyu Zhang and Prof. Daniela Damian, will share their expertise with researchers at Lero, enabling collaboration and knowledge sharing. The fellowship programme has brought world-class software researchers to Ireland to share their experience and expertise.
Researchers at Tel Aviv University developed a method to grow ultra-long and narrow graphene nanoribbons with semiconducting properties, opening doors for technological applications in advanced switching devices and spintronic systems. The study's success demonstrates a breakthrough in carbon-based nanomaterials.
Researchers at UMass Amherst have developed an analog computing device called a memristor that can complete complex scientific tasks while reducing energy consumption. The device uses physical laws to perform computations in a massively parallel fashion, accelerating matrix operations and overcoming the limitations of digital computing.
Newly designed analog chips combine digital and analog computing, providing high precision and low energy consumption. This innovation enables faster development of artificial intelligence (AI) systems and expands applications beyond traditional low-precision territory.
Researchers at Princeton University have developed a new AI chip that can run powerful AI systems using significantly less energy than existing semiconductors. The chip is designed to be compact and efficient, enabling its deployment in dynamic environments such as laptops, phones, and data centers.
Researchers have created a computer using an array of VCSELs that leverages optical feedback to efficiently solve complex optimization problems. The system encodes information in linear polarization states, minimizing interactions between variables and overcoming the von Neumann bottleneck.
Researchers have developed a novel optical neural network architecture that achieves nonlinear optical computation by precisely controlling ultrashort pulse propagation in multimode fibers. This approach streamlines the need for energy-intensive digital processes, achieving comparable accuracy with significantly reduced parameters.
Incheon National University researchers developed a web 3.0 streaming architecture that reduces delay, improves user experience, and ensures transparency and fairness for real-time services. The proposed system uses Inter-Planetary file system (IPFS) to enable blockchain-based peer-to-peer data storage and caching.
Researchers extend spatially incoherent diffractive networks to perform complex-valued linear transformations with negligible error, opening up new applications in fields like autonomous vehicles. This breakthrough enables the encryption and decryption of complex-valued images using spatially incoherent diffractive networks.
Researchers will incorporate advanced semiconductor technologies and AI into a millimeter-wave radio system to increase bandwidth while reducing energy consumption. The project aims to save tens to hundreds of terawatt-hours of energy per year, contributing to climate change mitigation.
A new study using twisted magnets as computational medium has made brain-inspired computing more adaptable, reducing energy use and potential carbon emissions. The research found that by applying magnetic fields and changing temperature, physical properties of the materials can be adapted to suit different machine-learning tasks.
Researchers developed a novel photonic processor with adaptive neural connectivity, allowing for the creation of complex artificial neural networks. The system utilizes waveguide-coupled phase-change material to create almost 8,400 optical neurons that can adapt their connections through synaptic and structural plasticity.
Researchers develop integrated photonic-electronic hardware capable of processing three-dimensional (3D) data, doubling parallelism for AI tasks and significantly boosting energy efficiency. The new chip can process 100 electrocardiogram signals simultaneously with high accuracy, outperforming electronic processors.
The High Performance Data Facility Hub will provide researchers with unprecedented data management resources, accelerating scientific discovery through seamless access to large and complex datasets. The hub will be led by Jefferson Lab and partner with Lawrence Berkeley National Laboratory.
Researchers aim to develop programmable formal specification-based data stream processor and hardware monitor to enhance microchip security and prevent malfunctions. They will explore novel technologies for real-time monitoring of physical and biological systems, including signal patterns within computer chips.