Training deep learning models typically requires a huge amount of labeled data which is expensive to acquire,especially in dense prediction tasks such as semantic ***,plant phenotyping datasets pose additional challen...
详细信息
Training deep learning models typically requires a huge amount of labeled data which is expensive to acquire,especially in dense prediction tasks such as semantic ***,plant phenotyping datasets pose additional challenges of heavy occlusion and varied lighting conditions which makes annotations more time-consuming to *** learning helps in reducing the annotation cost by selecting samples for labeling which are most informative to the model,thus improving model performance with fewer *** learning for semantic segmentation has been well studied on datasets such as PASCAL VOC and ***,its effectiveness on plant datasets has not received much *** bridge this gap,we empirically study and benchmark the effectiveness of four uncertainty-based active learning strategies on three natural plant organ segmentation *** also study their behaviour in response to variations in training configurations in terms of augmentations used,the scale of training images,active learning batch sizes,and train-validation set splits.
The development of automated high-throughput experimental platforms has enabled fast sampling of high-dimensional decision *** reach target properties efficiently,these platforms are increasingly paired with intellige...
详细信息
The development of automated high-throughput experimental platforms has enabled fast sampling of high-dimensional decision *** reach target properties efficiently,these platforms are increasingly paired with intelligent experimental ***,current optimizers show limitations in maintaining sufficient exploration/exploitation balance for problems dealing with multiple conflicting objectives and complex ***,we devise an Evolution-Guided Bayesian Optimization(EGBO)algorithm that integrates selection pressure in parallel with a q-Noisy Expected Hypervolume Improvement(qNEHVI)optimizer;this not only solves for the Pareto Front(PF)efficiently but also achieves better coverage of the PF while limiting sampling in the infeasible space.
Neural network watermarking is currently the mainstream deep neural network (DNN) model copyright protection method, and the embedding of the watermark is accomplished by changing the internal structure of the model o...
详细信息
The ability to identify walking conditions correctly is essential for both diagnosing and treating gait abnormalities. This study utilized machine learning algorithms to analyze multivariate gait data obtained from 10...
详细信息
The expansion of the Internet of Things (IoT) has made it possible for numerous widespread objects to connect to one another and communicate with one another, leading to unprecedented data releases. However, regardles...
详细信息
The main purpose of multimodal machine translation (MMT) is to improve the quality of translation results by taking the corresponding visual context as an additional input. Recently many studies in neural machine tran...
详细信息
The influence of melting-casting and heat treatment processes on the hardness of 2618 aluminum alloy was investigated by hardness testing. The results show that the hardness of 2618 alloy increases at first and then d...
详细信息
Online laptop sales are at an all-time high as a result of the pandemic. A laptop is a must-have for working from home, as well as e-learning and other activities. The buyer is aided in making a purchasing decision by...
详细信息
This paper is concerned with finite-time H_(∞) filtering for Markov jump systems with uniform quantization. The objective is to design quantized mode-dependent filters to ensure that the filtering error system is not...
详细信息
This paper is concerned with finite-time H_(∞) filtering for Markov jump systems with uniform quantization. The objective is to design quantized mode-dependent filters to ensure that the filtering error system is not only mean-square finite-time bounded but also has a prescribed finite-time H_(∞) performance. First, the case where the switching modes of the filter align with those of the MJS is considered. A numerically tractable filter design approach is proposed utilizing a mode-dependent Lyapunov function, Schur’s complement, and Dynkin’s formula. Then, the study is extended to a scenario where the switching modes of the filter can differ from those of the MJS. To address this situation, a mode-mismatched filter design approach is developed by leveraging a hidden Markov model to describe the asynchronous mode switching and the double expectation formula. Finally, a spring system model subject to a Markov chain is employed to validate the effectiveness of the quantized filter design approaches.
Dear Editor,Dummy attack(DA), a deep stealthy but impactful data integrity attack on power industrial control processes, is recently recognized as hiding the corrupted measurements in normal measurements. In this lett...
详细信息
Dear Editor,Dummy attack(DA), a deep stealthy but impactful data integrity attack on power industrial control processes, is recently recognized as hiding the corrupted measurements in normal measurements. In this letter, targeting a more practical case, we aim to detect the oneshot DA, with the purpose of revealing the DA once it is ***, we first formulate an optimization problem to generate one-shot DAs. Then, an unsupervised data-driven approach based on a modified local outlier factor(MLOF) is proposed to detect them.
暂无评论