Nanowerk August 17, 2022 Compute-in-memory (CIM) based on resistive random-access memory (RRAM) meets the energy demand on edge devices by performing AI computation directly within RRAM. Although efficiency, versatility and accuracy are all indispensable for broad adoption of the technology, the inter-related trade-offs among them cannot be addressed by isolated improvements on any single abstraction level of the design. By co-optimizing across all hierarchies of the design from algorithms and architecture to circuits and devices, a team of researchers in the US (Stanford University, UC San Diego, University of Notre Dame, Pittsburg University) has developed NeuRRAM—a RRAM-based CIM chip that […]
Category Archives: Computing
World’s first ultra-fast photonic computing processor using polarization
Phys.org June 15, 2022 While wavelength-selective systems have widely proliferated, polarization-addressable active photonics has not seen notable progress, primarily because tunable and polarization-selective nanostructures have been elusive. Researchers in the UK have introduced hybridized-active-dielectric (HAD) nanowires to achieve polarization-selective tunability. They demonstrated the ability to use polarization as a parameter to selectively modulate the conductance of individual nanowires within a multi-nanowire system. By using polarization as the tunable vector, they showed matrix-vector multiplication in a nanowire device configuration. According to the researchers while the HAD nanowires use phase-change materials as the active material, this concept can be generalized to other […]
Ultrafast ‘camera’ captures hidden behavior of potential ‘neuromorphic’ material
Phys.org May 9, 2022 A team of researchers in the US (Brookhaven National Laboratory, Duke University, Oak Ridge National Laboratory) captured the hidden trajectory of atomic motion of vanadium dioxide (VO2) as it transitioned from an insulator to a metal in response to a pulse of light. Vanadium dioxide exhibits an insulator-metal transition near room temperature in which a small voltage or current can produce a large change in resistivity with switching that can mimic the behavior of both neurons and synapses. Those are the signals produced by electrons scattering off the atoms of the vanadium dioxide sample as atoms […]
DNA computer using glass beads increases parallel processing power
Phys.org March 29, 2022 In general, DNA computation conducted in individual tubes is slow in generating chemical outputs in response to chemical inputs and requires fluorescence readout. Researchers at Emory University have introduced a new paradigm for DNA computation where the chemical input is processed and transduced into a mechanical output using dynamic DNA-based motors operating far from equilibrium. They applied DNA as a coating to extremely small glass beads. In practice, the glass beads either roll across the surface of a base of gold or hold steady, depending on how the DNA coating interacts with molecules that have been […]
Researchers use tiny magnetic swirls to generate true random numbers
Science Daily February 7, 2022 Researchers at Brown University have shown that the local dynamics of skyrmions, in contrast to the global dynamics of a skyrmion, can be introduced to provide effective functionalities for versatile computing. A single skyrmion interacting with local pinning centres under thermal effects can fluctuate in time and switch between a small-skyrmion and a large-skyrmion state, thereby serving as a robust true random number generator for probabilistic computing. Moreover, neighbouring skyrmions exhibit an anti-correlated coupling in their fluctuation dynamics. Both the switching probability and the dynamic coupling strength can be tuned by modifying the applied magnetic […]
Researcher develops new tool for understanding hard computational problems that appear intractable
Phys.org January 10, 2022 The problem of optimizing over random structures emerges in many areas of science and engineering, ranging from statistical physics to machine learning and artificial intelligence. For many such structures, finding optimal solutions by means of fast algorithms is not known and often believed impossible to solve. At the same time, the formal hardness of these problems in the form of the complexity-theoretic NP-hardness is lacking. Researchers at MIT describe a new approach for algorithmic intractability in random structures, which is based on the topological disconnectivity property of the set of pairwise distances of near-optimal solutions, called […]
Scientists develop the next generation of reservoir computing
Phys.org September 21, 2021 Reservoir computing is a machine learning algorithm developed in the early 2000s and used to solve the “hardest of the hard” computing problems. It requires very small training data sets, uses linear optimization, and thus requires minimal computing resources. It does that using an artificial neural network which is a black box. A team of researchers in the US (Ohio State University, industry, Clackson University) investigated the “black box” and found that the whole reservoir computing system could be greatly simplified, dramatically reducing the need for computing resources and saving significant time. They tested their concept […]
Harnessing the hum of fluorescent lights for more efficient computing
Science Daily May 12, 2021 Magnetostriction, which causes the buzz of fluorescent lights and electrical transformers, occurs when a change in the shape of the material causes a change in magnetic field. An international team of researchers (USA – University of Michigan, Cornell University, SUNY Buffalo, UC Berkeley, University of Wisconsin, Purdue University, Germany) has developed a material made of a combination of iron and gallium which has at least twice as magnetostrictive and far less costly than other materials in its class. By freezing the iron-gallium alloy and preventing it from forming an ordered structure they were able to […]
Harnessing socially-distant molecular interactions for future computing
Nanowerk February 15, 2021 Researchers in Australia studied the electronic properties of magnesium phthalocyanine (MgPc) sprinkled on a metal surface and demonstrated that the quantum mechanical properties of electrons within the molecules (energy and spatial distribution) are significantly affected by the presence of neighbouring molecules. This effect was observed for intermolecular separation distances of several nanometres. Quantitative analysis of the experimental results and theoretical modelling showed that this interaction was due to mixing between the quantum mechanical orbitals of neighbouring molecules. The molecular orbital mixing leads to significant changes in electron energies and electron distribution symmetries. The long range of […]
Designing customized “brains” for robots
MIT News January 21, 2021 In complex situations robots often do not move quickly because perceiving stimuli and calculating a response takes a lot of computation which limits reaction time. A team of researchers in the US (MIT, Harvard University) used robomorphic computing to bridge the mismatch between a robot’s “mind” and body. Their system creates a customized hardware design to best serve a particular robot’s computing needs. The user inputs the parameters of a robot, the system translates these physical properties into mathematical matrices. These matrices contain many zero values that roughly correspond to movements that are impossible given […]