NeuronFlow: a neuromorphic processor architecture for Live AI applications
Orlando Moreira, Amirreza Yousefzadeh, Fabian Chersi, Gokturk Cinserin, Rik-Jan Zwartenkot, Ajay Kapoor, Qiao Peng, Peter Kievits, Mina A. Khoei, Louis Rouillard, Aimee Ferouge, Jonathan Tapson, Ashoka Visweswara
Abstract
Neuronflow is a neuromorphic, many core, data flow architecture that exploits brain-inspired concepts to deliver a scalable event-based processing engine for neuron networks in Live AI applications. Its design is inspired by brain biology, but not necessarily biologically plausible. The main design goal is the exploitation of sparsity to dramatically reduce latency and power consumption as required by sensor processing at the Edge.
Topics & Concepts
Neuromorphic engineeringComputer scienceComputer architectureScalabilityExploitLatency (audio)ArchitectureMulti-core processorMicroarchitecturePower consumptionEmbedded systemArtificial intelligenceArtificial neural networkPower (physics)Parallel computingTelecommunicationsOperating systemComputer securityVisual artsArtPhysicsQuantum mechanicsAdvanced Memory and Neural ComputingNeuroscience and Neural EngineeringCCD and CMOS Imaging Sensors