Results 121 to 130 of about 10,284,296 (292)
Benchmarking Federated Few-Shot Learning for Video-Based Action Recognition
IEEE AccessFew-shot action recognition aims to train a model to classify actions in videos using only a few examples, known as “shots,” per action class.
Nguyen Anh Tu +5 more
doaj +1 more source
A Feature Generator for Few-Shot Learning
16 pages, Accepted to ACCV ...
Heethanjan Kanagalingam +5 more
openaire +3 more sources
Advanced Materials Interfaces, EarlyView.A VO2 device can act as a functional analog for behavioral timescale synaptic plasticity noted in the mouse brain by current pulsing. ABSTRACT Realizing biologically relevant plasticity rules in hardware is a central goal in neuromorphic computing.
Siddharth Kumar +3 more
wiley +1 more source
Two‐Way Shape Memory Polymer Composite Gripper for Adaptive Robotic Applications
Advanced Materials Technologies, EarlyView.A two‐way shape memory polymer (SMP) composite is developed with intrinsic shape‐changing capability driven solely by temperature, eliminating external actuation loads. Embedding the SMP in a low‐stiffness elastomeric matrix enabled reversible transformations during heating and cooling cycles.
Aamna Hameed, Kamran Ahmed Khan
wiley +1 more source
Advanced Materials Technologies, EarlyView.Desktop vacuum thermoforming enables rapid, low‐cost fabrication of 96‐channel polystyrene culture chips with very thin imaging interfaces. The resulting platform supports cell lines, primary cells, and patient‐derived organoids, and enables high‐resolution confocal microscopy and single‐cell visualization, providing an accessible approach for advanced
Aleksandra Fomina +5 more
wiley +1 more source
Few-shot Learning in Intelligent Agriculture: A Review of Methods and Applications
Journal of Agricultural SciencesDue to the high cost of data acquisition in many specific fields, such as intelligent agriculture, the available data is insufficient for the typical deep learning paradigm to show its superior performance.
Sezai Ercisli +7 more
doaj +1 more source
Prototypical Networks for Few-shot Learning
CoRR, 2017 We propose prototypical networks for the problem of few-shot classification, where a classifier must generalize to new classes not seen in the training set, given only a small number of examples of each new class. Prototypical networks learn a metric space in which classification can be performed by computing distances to prototype representations of ...
Jake Snell +2 more
openaire +4 more sources
Customizable Piezo‐Transmittance‐Based Multiaxial Force Sensor With Deep‐Learning‐Assisted Sensing
Advanced Materials Technologies, EarlyView.Development of a soft tri‐axis piezo‐transmittance sensor with integrated LEDs and photodiodes enables multidirectional force sensing through deformation‐induced optical modulation. A tailored 1D convolutional neural network provides real‐time force direction and intensity prediction, achieving 99.6% classification accuracy and a mean squared error of ...
Filippo Peruzzi +6 more
wiley +1 more source
Advanced Materials Technologies, EarlyView.Single‐filament material extrusion enables spatial stiffness programming through locally controlled thermal foaming of TPU. A data‐driven G‐code framework tunes porosity from 15.7% to 40.7% and hardness from 90 to 63 Shore A within and across layers. Monolithic hands and soft actuators consequently achieve distinct programmed motions without material ...
Rosanna Rifino +3 more
wiley +1 more source
Few-shot Calligraphy Style Learning
CoRRWe introduced "Presidifussion," a novel approach to learning and replicating the unique style of calligraphy of President Xu, using a pretrained diffusion model adapted through a two-stage training process. Initially, our model is pretrained on a diverse dataset containing works from various calligraphers.
Fangda Chen +3 more
openaire +3 more sources