Fig 3 - available via license: Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International
Content may be subject to copyright.
Source publication
There is an emerging need for predictive models to be trained on-the-fly, since in numerous machine learning applications data are arriving in an online fashion. A critical challenge encountered is that of limited availability of ground truth information (e.g., labels in classification tasks) as new data are observed one-by-one online, while anothe...
Contexts in source publication
Context 1
... examine now the memory size's role; we compare NN, and VGG with 3 blocks which achieved the best overall performance in Section VI-A. The budget is fixed to 0.1. In these experiments, we present the learning curves, that display the performance at different time steps. Fig. 3a shows ActiQ-NN's performance in MNIST with 10% imbalance. Figs. 3b and 3c show ActiQ-VGG's performance in MNIST with 10% and 1% imbalance respectively. Important remarks ...
Context 2
... examine now the memory size's role; we compare NN, and VGG with 3 blocks which achieved the best overall performance in Section VI-A. The budget is fixed to 0.1. In these experiments, we present the learning curves, that display the performance at different time steps. Fig. 3a shows ActiQ-NN's performance in MNIST with 10% imbalance. Figs. 3b and 3c show ActiQ-VGG's performance in MNIST with 10% and 1% imbalance respectively. Important remarks ...
Similar publications
The performance of deep learning-based detection methods has made them an attractive option for robotic perception. However, their training typically requires large volumes of data containing all the various situations the robots may potentially encounter during their routine operation. Thus, the workforce required for data collection and annotatio...
