Physics-driven human-like working memory outperforms digital networks in dynamic vision

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Liu, Jingli, Zheng, Huannan, Zou, Bohao, Yang, Kezhou
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866915890352619520
author Liu, Jingli
Zheng, Huannan
Zou, Bohao
Yang, Kezhou
author_facet Liu, Jingli
Zheng, Huannan
Zou, Bohao
Yang, Kezhou
contents While the unsustainable energy cost of artificial intelligence necessitates physics-driven computing, its performance superiority over full-precision GPUs remains a challenge. We bridge this gap by repurposing the Joule-heating relaxation dynamics of magnetic tunnel junctions, conventionally suppressed as noise, into neuronal intrinsic plasticity, realizing working memory with human-like features. Traditional AI utilizes energy-intensive digital memory that accumulates historical noise in dynamic environments. Conversely, our Intrinsic Plasticity Network (IPNet) leverages thermodynamic dissipation as a temporal filter. We provide direct system-level evidence that this physics-driven memory yields an 18x error reduction compared to spatiotemporal convolutional models in dynamic vision tasks, reducing memory-energy overhead by >90,000x. In autonomous driving, IPNet reduces prediction errors by 12.4% versus recurrent networks. This establishes a neuromorphic paradigm that shatters efficiency limits and surpasses conventional algorithmic performance.
format Preprint
id arxiv_https___arxiv_org_abs_2512_15829
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Physics-driven human-like working memory outperforms digital networks in dynamic vision
Liu, Jingli
Zheng, Huannan
Zou, Bohao
Yang, Kezhou
Emerging Technologies
Artificial Intelligence
Computer Vision and Pattern Recognition
Neural and Evolutionary Computing
While the unsustainable energy cost of artificial intelligence necessitates physics-driven computing, its performance superiority over full-precision GPUs remains a challenge. We bridge this gap by repurposing the Joule-heating relaxation dynamics of magnetic tunnel junctions, conventionally suppressed as noise, into neuronal intrinsic plasticity, realizing working memory with human-like features. Traditional AI utilizes energy-intensive digital memory that accumulates historical noise in dynamic environments. Conversely, our Intrinsic Plasticity Network (IPNet) leverages thermodynamic dissipation as a temporal filter. We provide direct system-level evidence that this physics-driven memory yields an 18x error reduction compared to spatiotemporal convolutional models in dynamic vision tasks, reducing memory-energy overhead by >90,000x. In autonomous driving, IPNet reduces prediction errors by 12.4% versus recurrent networks. This establishes a neuromorphic paradigm that shatters efficiency limits and surpasses conventional algorithmic performance.
title Physics-driven human-like working memory outperforms digital networks in dynamic vision
topic Emerging Technologies
Artificial Intelligence
Computer Vision and Pattern Recognition
Neural and Evolutionary Computing
url https://arxiv.org/abs/2512.15829