A5064191197- Advanced Memory and Neural Computing
- CCD and CMOS Imaging Sensors
- Neural dynamics and brain function
- Neural Networks and Applications
- Ferroelectric and Negative Capacitance Devices
- Neuroscience and Neural Engineering
- Photoreceptor and optogenetics research
- Neural Networks and Reservoir Computing
- Fault Detection and Control Systems
- Software System Performance and Reliability
- Distributed and Parallel Computing Systems
- Blind Source Separation Techniques
- Wireless Signal Modulation Classification
- Visual Attention and Saliency Detection
- Image and Signal Denoising Methods
- Modular Robots and Swarm Intelligence
- Radar Systems and Signal Processing
- Parallel Computing and Optimization Techniques
- Optical Systems and Laser Technology
- Cloud Computing and Resource Management
IBM (United States)
2013-2023
IBM Research - Almaden
2012-2016
The new era of cognitive computing brings forth the grand challenge developing systems capable processing massive amounts noisy multisensory data. This type intelligent poses a set constraints, including real-time operation, low-power consumption and scalability, which require radical departure from conventional system design. Brain-inspired architectures offer tremendous promise in this area. To end, we developed TrueNorth, 65 mW neurosynaptic processor that implements non-von Neumann,...
Significance Brain-inspired computing seeks to develop new technologies that solve real-world problems while remaining grounded in the physical requirements of energy, speed, and size. Meeting these challenges requires high-performing algorithms are capable running on efficient hardware. Here, we adapt deep convolutional neural networks, which today’s state-of-the-art approach for machine perception many domains, perform classification tasks neuromorphic hardware, is most platform networks....
Marching along the DARPA SyNAPSE roadmap, IBM unveils a trilogy of innovations towards TrueNorth cognitive computing system inspired by brain's function and efficiency. Judiciously balancing dual objectives functional capability implementation/operational cost, we develop simple, digital, reconfigurable, versatile spiking neuron model that supports one-to-one equivalence between hardware simulation is implementable using only 1272 ASIC gates. Starting with classic leaky integrate-and-fire...
IBM's brain-inspired processor is a massively parallel neural network inference engine containing 1 million spiking neurons and 256 low-precision synapses. Now, after decade of fundamental research spanning neuroscience, architecture, chips, systems, software, algorithms, IBM has delivered the largest neurosynaptic computer ever built.
Marching along the DARPA SyNAPSE roadmap, IBM unveils a trilogy of innovations towards TrueNorth cognitive computing system inspired by brain's function and efficiency. The sequential programming paradigm von Neumann architecture is wholly unsuited for TrueNorth. Therefore, as our main contribution, we develop new that permits construction complex algorithms applications while being efficient effective programmer productivity. consists (a) an abstraction program, named Corelet, representing...
Marching along the DARPA SyNAPSE roadmap, IBM unveils a trilogy of innovations towards TrueNorth cognitive computing system inspired by brain's function and efficiency. The non-von Neumann nature architecture necessitates novel approach to efficient design. To this end, we have developed set abstractions, algorithms, applications that are natively for TrueNorth. First, repeatedly-used abstractions span neural codes (such as binary, rate, population, time-to-spike), long-range connectivity,...
Computing, since its inception, has been processor-centric, with memory separated from compute. Inspired by the organic brain and optimized for inorganic silicon, NorthPole is a neural inference architecture that blurs this boundary eliminating off-chip memory, intertwining compute on-chip, appearing externally as an active chip. low-precision, massively parallel, densely interconnected, energy-efficient, spatial computing co-optimized, high-utilization programming model. On ResNet50...
Inspired by the function, power, and volume of organic brain, we are developing TrueNorth, a novel modular, non-von Neumann, ultra-low compact architecture. TrueNorth consists scalable network neurosynaptic cores, with each core containing neurons, dendrites, synapses, axons. To set sail for developed Compass, multi-threaded, massively parallel functional simulator compiler that maps long-distance pathways in macaque monkey brain to TrueNorth. We demonstrate near-perfect weak scaling on 16...
Inspired by the function, power, and volume of organic brain, we are developing TrueNorth, a novel modular, non-von Neumann, ultra-low compact architecture. TrueNorth consists scalable network neurosynaptic cores, with each core containing neurons, dendrites, synapses, axons. To set sail for developed Compass, multi-threaded, massively parallel functional simulator compiler that maps long-distance pathways in macaque monkey brain to TrueNorth. We demonstrate near-perfect weak scaling on 16...
Drawing on neuroscience, we have developed a parallel, event-driven kernel for neurosynaptic computation, that is efficient with respect to memory, and communication. Building the previously demonstrated highly optimized software expression of kernel, here, demonstrate True North, co-designed silicon kernel. North achieves five orders magnitude reduction in energy to-solution two speedup time-to solution, when running computer vision applications complex recurrent neural network simulations....
This paper describes the hardware and software ecosystem encompassing brain-inspired TrueNorth processor - a 70mW reconfigurable silicon chip with 1 million neurons, 256 synapses, 4096 parallel distributed neural cores. For systems, we present scale-out system loosely coupling 16 single-chip boards scale-up tightly integrating chips in 4 × configuration by exploiting TrueNorth's native tiling. software, an end-to-end consisting of simulator, programming language, integrated environment,...
The Deep Neural Network (DNN) era was ushered in by the triad of algorithms, big data, and more powerful hardware processors for training large-scale neural networks. Now, ubiquitous deployment DNNs inference edge, embedded, data center applications demands power-efficient processors, while attaining increasingly higher computational performance. To address this Inference Challenge, we developed NorthPole Architecture implemented a Chip instantiation [1, 2].
Identifying interesting or salient regions in an image plays important role for multimedia search, object tracking, active vision, segmentation, and classification. Existing saliency extraction algorithms are implemented using the conventional von Neumann computational model. We propose a bottom-up model of visual saliency, inspired by primate cortex, which is compatible with TrueNorth-a low-power, brain-inspired neuromorphic substrate that runs large-scale spiking neural networks real-time....
Resource requirement prediction of jobs prior to their submission in the dynamic and heterogeneous natured modern distributed systems is a difficult task. A feedback-based job modelling scheme based on clone detection technique was proposed [1, 2] an extended taxonomy clones facilitate [3]. In this paper, we propose resource for different Near Miss clones. The differ only some aspects like data type input data, size operations performed job. So, new from traits its previously executed system...