Results 71 to 80 of about 8,974 (165)
Sign-MAML: Efficient Model-Agnostic Meta-Learning by SignSGD
, 2021 We propose a new computationally-efficient first-order algorithm for Model-Agnostic Meta-Learning (MAML). The key enabling technique is to interpret MAML as a bilevel optimization (BLO) problem and leverage the sign-based SGD(signSGD) as a lower-level optimizer of BLO.
Fan, Chen, Ram, Parikshit, Liu, Sijia
openaire +2 more sources
Unsupervised Online Learning for Substation Equipment Defect Identification Based on SoftHebb
Energy Internet, Volume 2, Issue 4, Page 287-296, December 2025.ABSTRACT With the advancement of artificial intelligence, deep learning algorithms have increasingly been adopted for defect detection in substation equipment. However, a widely recognised limitation of such models is their inability to learn autonomously over time, resulting in performance degradation when faced with changing data distributions—a ...
Zhisong Zhang +6 more
wiley +1 more source
, 2019 The field of few-shot learning has recently seen substantial advancements. Most of these advancements came from casting few-shot learning as a meta-learning problem. Model Agnostic Meta Learning or MAML is currently one of the best approaches for few-shot learning via meta-learning.
Antoniou, Antreas +2 more
openaire
La-MAML: Look-ahead Meta Learning for Continual Learning
, 2020 The continual learning problem involves training models with limited capacity to perform well on a set of an unknown number of sequentially arriving tasks. While meta-learning shows great potential for reducing interference between old and new tasks, the current training procedures tend to be either slow or offline, and sensitive to many hyper ...
Gupta, Gunshi +2 more
openaire +2 more sources
Computer-Aided Civil and Infrastructure Engineering, Volume 40, Issue 29, Page 5512-5533, 9 December 2025.Abstract Rapid and accurate estimation of seismic responses in city‐scale buildings is critical for post‐earthquake loss assessment and pre‐event identification of vulnerable buildings. However, conventional numerical simulation methods struggle to balance efficiency and accuracy when applied to large‐scale buildings, while existing data‐driven methods
Chenyu Zhang +3 more
wiley +1 more source
How Does the Task Landscape Affect MAML Performance?
, 2020 Model-Agnostic Meta-Learning (MAML) has become increasingly popular for training models that can quickly adapt to new tasks via one or few stochastic gradient descent steps. However, the MAML objective is significantly more difficult to optimize compared to standard non-adaptive learning (NAL), and little is understood about how much MAML improves over
Collins, Liam +2 more
openaire +2 more sources
Fixed-MAML for Few-shot Classification in Multilingual Speech Emotion Recognition
, 2022 In this paper, we analyze the feasibility of applying few-shot learning to speech emotion recognition task (SER). The current speech emotion recognition models work exceptionally well but fail when then input is multilingual. Moreover, when training such models, the models' performance is suitable only when the training corpus is vast.
Naman, Anugunj +2 more
openaire +2 more sources
Computer-Aided Civil and Infrastructure Engineering, Volume 40, Issue 28, Page 5212-5227, 28 November 2025.Abstract This study introduces a novel integrated framework for real‐time tunnel boring machine (TBM) thrust prediction, addressing critical limitations in handling non‐stationarity, complex spatiotemporal dependencies, and abrupt disturbances. First, a real‐time windowed multi‐resolution analysis process, which performs decomposition strictly within ...
Yongxin Wu +4 more
wiley +1 more source
Tibetan Few‐Shot Learning Model With Deep Contextualised Two‐Level Word Embeddings
CAAI Transactions on Intelligence Technology, Volume 10, Issue 5, Page 1394-1410, October 2025.ABSTRACT Few‐shot learning is the task of identifying new text categories from a limited set of training examples. The two key challenges in few‐shot learning are insufficient understanding of new samples and imperfect modelling. The uniqueness of low‐resource languages lies in their limited linguistic resources, which directly leads to the difficulty ...
Ziyue Zhang +11 more
wiley +1 more source
ES-MAML: Simple Hessian-Free Meta Learning
, 2019 We introduce ES-MAML, a new framework for solving the model agnostic meta learning (MAML) problem based on Evolution Strategies (ES). Existing algorithms for MAML are based on policy gradients, and incur significant difficulties when attempting to estimate second derivatives using backpropagation on stochastic policies. We show how ES can be applied to
Song, Xingyou +5 more
openaire +2 more sources