AI Glossary
The complete dictionary of Artificial Intelligence
Spiking Neural Networks
Artificial neural networks that communicate through discrete timed signals (spikes) rather than continuous values, more faithfully mimicking the functioning of the biological brain. These architectures enable more energy-efficient information processing.
Memristor
Non-linear passive electronic component whose resistance depends on the history of the current that has passed through it, simulating the behavior of a biological synapse. Memristors are fundamental for materially implementing synaptic plasticity in neuromorphic architectures.
Synaptic Plasticity
Mechanism by which the strength of synaptic connections between neurons can be modified based on activity, forming the basis of learning and memory. In neuromorphic computing, it is implemented materialistically to enable autonomous adaptation of circuits.
Neuromorphic Chip
Integrated circuit specially designed to implement neuromorphic computing architectures with physical neurons and synapses. These chips optimize parallel processing and low energy consumption typical of biological systems.
Crossbar Array
Dense matrix interconnection structure enabling efficient connection of numerous neurons and synapses in neuromorphic architectures. This organization allows compact and energy-efficient hardware implementation of massively parallel neural networks.
Analog Computing
Computation method using continuous signals to represent and process information, offering superior energy efficiency for certain neuromorphic operations. This approach enables more natural implementation of biological processes like temporal integration.
Hardware Acceleration
Use of specialized circuits to accelerate the execution of neuromorphic computing tasks, surpassing the performance of general-purpose processors. In this context, it enables real-time implementation of complex neural networks.
Neuro-inspired Architecture
System design inspired by the organizational principles of the brain without necessarily faithfully reproducing them, optimizing performance for specific applications. These architectures balance biological fidelity and computational efficiency.
Brain-Inspired Computing
Interdisciplinary field aiming to develop computing systems based on the brain's operating principles, including massive parallel processing and adaptive learning. This approach promises breakthroughs in artificial intelligence and energy efficiency.
Low-Power Computing
Optimization of computing systems to minimize their energy consumption, a key objective of neuromorphic computing inspired by the remarkable efficiency of the human brain. Neuromorphic architectures achieve unparalleled levels of energy efficiency.
Neural Core
Fundamental computing unit in neuromorphic processors containing all circuits necessary for emulating a group of neurons and their connections. Each core operates semi-autonomously to allow massive scalability.
Resistive RAM (RRAM)
Type of non-volatile memory whose resistance can be modified in a controlled manner, ideal for implementing plastic synapses in neuromorphic circuits. These devices enable dense integration and low power consumption.