Results 171 to 180 of about 3,000 (250)
Stability-Controlled Continual Federated Learning for Energy-Harvesting AIoT Systems. [PDF]
Park J, Yoon I, Noh DK.
europepmc +1 more source
ABSTRACT This systematic review synthesizes evidence from 68 studies, including peer‐reviewed journal articles, indexed conference/workshop proceedings and five remaining arXiv preprints published between 2022 and 2025, on small language models (SLMs) as computationally efficient alternatives to large language models (LLMs).
Sena Dikici, Turgay Tugay Bilgin
wiley +1 more source
Evolutionary Digital Twin for Oil and Gas Pipelines: A Cognitive Multi-Agent Framework with Continuous Feedback Learning. [PDF]
Shi N +7 more
europepmc +1 more source
Powerful representation of the poor? German welfare associations' narrative advocacy during COVID‐19
Abstract The COVID‐19 pandemic sparked unprecedented experimentation in the German social assistance system, leading to changes previously considered impracticable by policymakers. This included a sanctions moratorium, easier access to benefits, and temporary cash transfers, all of which were advocated by welfare associations—key organized interests ...
Christopher Smith Ochoa
wiley +1 more source
Spiking Neural Networks with Continual Learning for Steering Angle Regression: A Sustainable AI Perspective. [PDF]
Martínez FS, Costa S, Parada R.
europepmc +1 more source
A multiagent MASAC strategy for Volt‐VAR control with grid‐forming storage. Dynamic zoning and a multi‐timescale replay mechanism enhance adaptability. A safety filter prevents voltage violations during learning and operation. Achieves lowest voltage deviation and fast recovery in volatile conditions.
Kai Guan +5 more
wiley +1 more source
CNN‐DDQN Spatial Learning for Fractional Frequency Reuse Optimization in 5G Heterogeneous Networks
This study proposes a spatially aware DRL framework that uses multichannel network‐state tensors and a lightweight CNN‐DDQN architecture to capture interference patterns and user clustering in 5G heterogeneous networks (HetNets). The model is computationally efficient with only 7361 trainable parameters, achieving 115× faster execution and < 10 ms ...
Christian Onu +9 more
wiley +1 more source
ABSTRACT The concept of sustainability has been shaped by the history of environmental problems and ecological crises, and the scientific value of studying the past has long been recognized. However, the relevance of studying Earth's past in the context of the sustainability debate goes beyond testing Earth‐system models used for future climate ...
Georg Feulner
wiley +1 more source
Gated subspace alignment with drift compensation for parameter-efficient Class-Incremental Learning. [PDF]
Gu J, Huang S, Li T, Zhang S, Li M.
europepmc +1 more source
I'll Be Dammed: Efficiency and Cost‐Analysis of a Rehabilitated Prehispanic Dam in the Andes
ABSTRACT Within archaeology there is a growing shift away from the purely aesthetic towards making the past relevant to present and future needs and demands. The Past Water Futures project charts the impact of ancient dam rehabilitation on climate change‐induced water scarcity in the North Central Andes.
Kevin Lane +5 more
wiley +1 more source

