Category Ranking
98%
Total Visits
921
Avg Visit Duration
2 minutes
Citations
20
Article Abstract
The brain is dynamic, associative, and efficient. It reconfigures by associating the inputs with past experiences, with fused memory and processing. In contrast, AI models are static, unable to associate inputs with past experiences, and run on digital computers with physically separated memory and processing. We propose a hardware-software co-design, a semantic memory-based dynamic neural network using a memristor. The network associates incoming data with the past experience stored as semantic vectors. The network and the semantic memory are physically implemented on noise-robust ternary memristor-based computing-in-memory (CIM) and content-addressable memory (CAM) circuits, respectively. We validate our co-designs, using a 40-nm memristor macro, on ResNet and PointNet++ for classifying images and three-dimensional points from the MNIST and ModelNet datasets, which achieves not only accuracy on par with software but also a 48.1 and 15.9% reduction in computational budget. Moreover, it delivers a 77.6 and 93.3% reduction in energy consumption.
Download full-text PDF |
Source |
|---|---|
| http://www.ncbi.nlm.nih.gov/pmc/articles/PMC11323881 | PMC |
| http://dx.doi.org/10.1126/sciadv.ado1058 | DOI Listing |
Publication Analysis
Top Keywords
Similar Publications
Attention-aware semantic relevance predicting Chinese sentence reading.
Cognition
February 2025
College of Foreign Languages and Literature, Fudan University, China.
In recent years, several influential computational models and metrics have been proposed to predict how humans comprehend and process sentence. One particularly promising approach is contextual semantic similarity. Inspired by the attention algorithm in Transformer and human memory mechanisms, this study proposes an "attention-aware" approach for computing contextual semantic relevance.
View Article and Find Full Text PDFTransitional gradation and the distinction between episodic and semantic memory.
Philos Trans R Soc Lond B Biol Sci
November 2024
Philosophy, University of Florida, Gainesville, FL 32611, USA.
In this article, we explore various arguments against the traditional distinction between episodic and semantic memory based on the metaphysical phenomenon of transitional gradation. Transitional gradation occurs when two candidate kinds A and B grade into one another along a continuum according to their characteristic properties. We review two kinds of arguments-from the gradual semanticization of episodic memories as they are consolidated, and from the composition of episodic memories during storage and recall from semantic memories-that predict the proliferation of such transitional forms.
View Article and Find Full Text PDFEvent boundaries help structure the content of episodic memories by segmenting continuous experiences into discrete events. Event boundaries may also serve to preserve meaningful information within an event, thereby actively separating important memories from interfering representations imposed by past and future events. Here, we tested the hypothesis that event boundaries organize emotional memory based on changing dynamics as events unfold.
View Article and Find Full Text PDFSemantic memory-based dynamic neural network using memristive ternary CIM and CAM for 2D and 3D vision.
Sci Adv
August 2024
Key Laboratory of Microelectronic Devices and Integrated Technology, Institute of Microelectronics, Chinese Academy of Sciences, Beijing 100049, China.
Article Synopsis
- - The brain functions dynamically by reconfiguring itself to associate new inputs with past experiences, while AI models are static and don't have this ability, using separate memory and processing systems.
- - The authors propose a new approach combining hardware and software in a dynamic neural network that uses memristors to create a semantic memory system, allowing for the association of new data with past experiences.
- - Their designs, tested on ResNet and PointNet++ for image and 3D point classification, show high accuracy comparable to traditional software methods, and result in significant reductions in both computational budget and energy consumption.
Default mode network shows distinct emotional and contextual responses yet common effects of retrieval demands across tasks.
Hum Brain Mapp
May 2024
Department of Psychology, University of York, York, UK.
The default mode network (DMN) lies towards the heteromodal end of the principal gradient of intrinsic connectivity, maximally separated from the sensory-motor cortex. It supports memory-based cognition, including the capacity to retrieve conceptual and evaluative information from sensory inputs, and to generate meaningful states internally; however, the functional organisation of DMN that can support these distinct modes of retrieval remains unclear. We used fMRI to examine whether activation within subsystems of DMN differed as a function of retrieval demands, or the type of association to be retrieved, or both.
View Article and Find Full Text PDF