Exploiting neuro-inspired dynamic sparsity for energy-efficient intelligent perception

Artificial intelligence (AI) has made significant strides towards efficient online processing of sensory signals at the edge through the use of deep neural networks with ever-expanding size. However, this trend has brought with it escalating computational costs and energy consumption, which have bec...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Veröffentlicht in:Nature communications Jg. 16; H. 1; S. 9928 - 15
Hauptverfasser: Zhou, Sheng, Gao, Chang, Delbruck, Tobi, Verhelst, Marian, Liu, Shih-Chii
Format: Journal Article
Sprache:Englisch
Veröffentlicht: London Nature Publishing Group UK 11.11.2025
Nature Publishing Group
Nature Portfolio
Schlagworte:
639/166
639/166/987
Artificial intelligence
Artificial neural networks
Brain
Co-design
Computing costs
Energy consumption
Energy costs
Energy efficiency
Humanities and Social Sciences
Information processing
Memory
multidisciplinary
Neural networks
Neurons
Optimization techniques
Perception
Perspective
Science
Science (multidisciplinary)
Sensors
Sensory integration
Signal processing
Sparsity
ISSN:2041-1723, 2041-1723
Online-Zugang:Volltext
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
Beschreibung
Zusammenfassung:Artificial intelligence (AI) has made significant strides towards efficient online processing of sensory signals at the edge through the use of deep neural networks with ever-expanding size. However, this trend has brought with it escalating computational costs and energy consumption, which have become major obstacles to the deployment and further upscaling of these models. In this Perspective, we present a neuro-inspired vision to boost the energy efficiency of AI for perception by leveraging brain-like dynamic sparsity. We categorize various forms of dynamic sparsity rooted in data redundancy and discuss potential strategies to enhance and exploit it through algorithm-hardware co-design. Additionally, we explore the technological, architectural, and algorithmic challenges that need to be addressed to fully unlock the potential of dynamic-sparsity-aware neuro-inspired AI for energy-efficient perception. Edge AI enables intelligent perception in sensory devices, yet at excessive energy costs. This Perspective outlines a neuro-inspired vision for efficient edge perception, sketching the design space of data-driven and stateful dynamic sparsity to selectively activate sensors, memory, and compute.
Bibliographie:ObjectType-Article-1
SourceType-Scholarly Journals-1
ObjectType-Feature-2
content type line 14
ObjectType-Review-3
content type line 23
ISSN:2041-1723
2041-1723
DOI:10.1038/s41467-025-65387-7