Enterprise Technology Glossary

6,173 results · page 183 of 309

• Neural Inference Accelerator

  Neural inference accelerator is a hardware component or subsystem that runs trained neural network models for inference workloads with higher efficiency than general-purpose processors, which matters to enterprises optimizing performance, latency, and power use for production AI applications at scale.

• Neural Network

  Neural network is a computational model built from interconnected layers of artificial neurons that learn patterns or mappings from data by adjusting numerical parameters. It matters in enterprises because it underlies many machine learning applications used in analytics, automation, and decision support.

• Neural Network Accelerator

  Neural network accelerator is a specialized hardware component that executes neural network computations more efficiently than general-purpose processors, enabling enterprises to run AI inference and training workloads with higher throughput, lower latency, and improved resource utilization in data center, edge, and embedded environments.

• Neural Network Backdoor

  Neural network backdoor is a hidden, attacker-inserted behavior in a trained model that activates only under specific triggers, creating targeted misclassifications or outcomes and posing integrity, safety, and supply chain risk for enterprise AI systems and applications.

• Neural Network Processor

  Neural network processor is a specialized compute unit that executes neural network and deep learning workloads using parallel tensor or matrix operations, enabling enterprises to run AI inference and, in some cases training, within specific performance, latency, and power constraints.

• Neural Network Pruning

  Neural network pruning is a model compression technique that removes parameters or structures from trained neural networks to reduce compute, memory, and energy usage while keeping accuracy within defined tolerances, supporting deployment on constrained hardware and more efficient enterprise AI operations.

• Neural Network Training

  Neural network training is the process of iteratively adjusting a neural network’s parameters using optimization algorithms and loss functions so the model approximates desired outputs, which matters for enterprises that build, govern, and operate AI and machine learning capabilities at scale.

• Neural Plasticity Engine

  Neural Plasticity Engine is not an established or standardized term in current academic, standards, or enterprise technology literature, and no vetted sources define its technical characteristics, architectural role, or business relevance as a distinct, recognized technology category.

• Neural Processing Unit

  Neural processing unit (NPU) is a specialized hardware accelerator for executing artificial neural network workloads, used in data centers, edge systems, and devices to increase performance per watt and throughput for AI inference compared with general-purpose compute architectures.

• Neural Radiance Field

  Neural radiance field (NeRF) is a neural network–based 3D scene representation that models color and volumetric density as continuous functions, enabling novel view synthesis from images. It matters for enterprises that need photorealistic 3D reconstructions for visualization, simulation, and digital twin workflows.

• Neuromorphic

  Neuromorphic computing is hardware and system design that emulates biological neural processing to execute event-driven, parallel, and energy-aware computation. It matters in enterprise contexts for low-power, low-latency AI, sensor analytics, and edge workloads that complement conventional CPUs, GPUs, and AI accelerators.

• Neuromorphic Accelerator

  Neuromorphic accelerator is a specialized processor that implements brain-inspired or spiking neural network models using neuron- and synapse-like circuits. It matters to enterprises that need low-power, always-on AI processing for edge, embedded, or sensor-centric workloads in constrained environments.

• Neuromorphic Chip

  Neuromorphic chip is a specialized processor that implements brain-inspired spiking or event-driven computation for machine learning and perception tasks, relevant to enterprises that require low-power, low-latency AI acceleration in edge, embedded, or heterogeneous data center architectures.

• Neuromorphic Computing

  Neuromorphic computing is a hardware and software approach that emulates biological neural systems using event-driven, parallel architectures to process information with low power and latency, relevant for enterprises exploring energy-efficient AI workloads in edge, embedded, and sensor-driven environments.

• Neuromorphic Processor

  Neuromorphic processor refers to specialized hardware that implements neuron- and synapse-inspired circuits for event-driven, parallel computation with constrained power and latency. It matters in enterprise contexts for handling perception and pattern-recognition workloads, particularly in edge, embedded, and sensor-centric systems where conventional accelerators face resource limits.

• Neuromorphic Sensor

  Neuromorphic sensor is a sensing device that uses neuromorphic engineering principles to output asynchronous spike or event streams instead of frames, enabling low-power, low-latency perception and on-sensor processing for enterprise edge, embedded, and autonomous system workloads.

• Neuro-Symbolic Integration

  Neuro-symbolic integration is an artificial intelligence approach that combines neural network–based learning with symbolic reasoning and knowledge representation, enabling enterprises to link data-driven models with explicit rules, ontologies, and knowledge graphs for governed decision support and alignment with domain and regulatory logic.

• Neuro-Symbolic Simulation

  Neuro-symbolic simulation is a computational approach that combines neural networks with symbolic reasoning or rule-based models to simulate complex systems under explicit constraints, supporting enterprises that require data-driven behavior aligned with formal rules, domain knowledge, compliance requirements, or structured decision logic.

• Neurosynaptic Core

  Neurosynaptic core is a neuromorphic computing building block that integrates neuron and synapse circuits with local memory to execute spiking neural networks in an event-driven, parallel manner, supporting low-power, real-time AI workloads in edge, embedded, and research-oriented enterprise systems.

• Next-Gen Cooling Readiness

  Next-Gen Cooling Readiness is the capability of a data center or digital infrastructure to support advanced cooling technologies such as liquid and immersion cooling, enabling higher-density compute deployments while remaining within technical, regulatory, efficiency, and operational reliability constraints.