Various approaches have been developed to upper bound the generalization error of a supervised learning algorithm. However, existing bounds are often loose and lack of guarantees. As a result, they may fail to charact...
详细信息
Aiming at the problem of ecological environment monitoring, the paper proposes an IoT technology for ecological environment monitoring based on wavelet sampling algorithm and multi-sensor network platform. This techno...
详细信息
In the context of large-scale grid connection of distributed energy, during the reconfiguration of the distribution network, the availability of distributed energy and the load of the distribution system may be incons...
详细信息
In the context of large-scale grid connection of distributed energy, during the reconfiguration of the distribution network, the availability of distributed energy and the load of the distribution system may be inconsistent with the prediction due to the influence of environmental factors and human factors. If the distribution network reconfiguration is still carried out according to the expected offline optimization scheme, there may be reliability problems of voltage over-limits and economic problems of increased network loss in the actual reconfiguration process. Therefore, the reconfiguration plan formulated in advance can give some guidance to the dispatch operator, but it may not be directly used in the actual reconfiguration process. This paper proposes a deep reinforcement learning approach to solving the electric distribution network reconfiguration. Based on the uncertainty of distributed energy output and network load in the distribution network, the online algorithm of distribution network reconfiguration realizes the second-level solution of distribution network reconfiguration, through day-ahead training of the neural network.
With the rapid development of intelligent networked vehicles and driverless technology, the importance of the dialogue between human and vehicle artificial intelligence has also become prominent. In the case of the dr...
详细信息
In collaboration with Amberg Technology, 1Institute of Rock Mechanics and Tunnelling at Graz University of Technology is developing a model to predict geological conditions ahead of the tunnel face. The model employs ...
详细信息
Alzheimer’s disease (AD) has become a major health problem over the past few decades. AD can be defined as a neurodegenerative disorder that causes the brain cells to degenerate and die. AD is the most popular cause ...
详细信息
Training multiple tasks jointly in one deep network yields reduced latency during inference and better performance over the single-task counterpart by sharing certain layers of a network. However, over-sharing a netwo...
详细信息
ISBN:
(纸本)9781713821120
Training multiple tasks jointly in one deep network yields reduced latency during inference and better performance over the single-task counterpart by sharing certain layers of a network. However, over-sharing a network could erroneously enforce over-generalization, causing negative knowledge transfer across tasks. Prior works rely on human intuition or pre-computed task relatedness scores for ad hoc branching structures. They provide sub-optimal end results and often require huge efforts for the trial-and-error process. In this work, we present an automated multi-task learning algorithm that learns where to share or branch within a network, designing an effective network topology that is directly optimized for multiple objectives across tasks. Specifically, we propose a novel tree-structured design space that casts a tree branching operation as a gumbel-softmax sampling procedure. This enables differentiable network splitting that is end-to-end trainable. We validate the proposed method on controlled synthetic data, CelebA, and Taskonomy.
It may be expected that survival stresses lead to evolutionary strengthening of the population. By eliminating the weakest individuals, stress can lead to increase in strength of the average individual. However, this ...
详细信息
As most of the fraud detection systems take part on the side of banking sector, aviation industry is one of the important sectors which faces seriously damaging fraud cases. As fraud cases increase exponentially, airl...
详细信息
In recent years, the development of deep learning has brought new ideas to internal threat detection. In this paper, three common deep learning algorithms for threat detection are optimized and innovated, and feature ...
详细信息
暂无评论