A Survey of Deep Active Learning

TL;DR: This study proposes AdaDS, a framework that adaptively selects data for knowledge distillation, leveraging various strategies to improve performance across tasks, data sizes, and training stages, achieving comparable results to DistilBERT with reduced computational cost.

TL;DR: This paper investigates the feasibility of implementing on-premise AIOps solutions by leveraging open-source tools and introduces a comprehensive A IOps infrastructure that has been successfully deployed in a company and provides the rationale behind different choices that were made to build its various components.

TL;DR: This study extends previous work by introducing a batch-mode deep active learning approach tailored for regression in graph-structured data, and deploys an array of base kernels, kernel transformations, and selection methods, informed by both Bayesian and non-Bayesian strategies.

TL;DR: In this article, the authors investigate the feasibility of performing semi-automatic annotation of video data using linear interpolation and show that it is possible to maintain the performance while decreasing the annotation workload by about 95%.

TL;DR: Neber et al. as discussed by the authors proposed a network and training strategy that relies on the strong use of data augmentation to use the available annotated samples more efficiently, which can be trained end-to-end from very few images and outperforms the prior best method (a sliding-window convolutional network) on the ISBI challenge for segmentation of neuronal structures in electron microscopic stacks.

TL;DR: The state-of-the-art performance of CNNs was achieved by Deep Convolutional Neural Networks (DCNNs) as discussed by the authors, which consists of five convolutional layers, some of which are followed by max-pooling layers, and three fully-connected layers with a final 1000-way softmax.

TL;DR: In this article, a graph transformer network (GTN) is proposed for handwritten character recognition, which can be used to synthesize a complex decision surface that can classify high-dimensional patterns, such as handwritten characters.

TL;DR: The key insight is to build “fully convolutional” networks that take input of arbitrary size and produce correspondingly-sized output with efficient inference and learning.