However, the previous work set fixed-length input in their models. In this way, they ignore shorter pedestrian trajectories. This approach leads to insufficient feature information and inaccurately prediction in some scenarios. In this paper, we propose an A utoencoder-based model for pede s ...
Experimental results on standard datasets indicate that AUCH achieves competitive results compared to state-of-the-art models for retrieval and clustering tasks.doi:10.1007/s10489-020-01797-yBolin ZhangJiangbo QianApplied Intelligence
Data from the same source holds similarity as the LUS image quality varies depending on the type of apparatus used to examine by the clinician, which enables the machine learning or deep learning models to predict with greater accuracy. One example is presented in Table 4 to check the ...
The use of latent factor models like those based on the autoencoders18 have been rising, stemming from the recent successes of (deep) neural network models for vision and speech tasks. Justifying their popularity in the recent years autoencoder based matrix imputation methods outperform the ...
detection, with the inclusion of an AL framework to balance the labeling budget throughout the data stream and test if this can further enhance performance by highlighting more relevant samples and using active change detection for both training new AE models, and selecting suitable threshold values...
The CUDA Toolkit53 is used to perform CAE models on the GPU and simultaneously test the latency and power consumption. The power statistics, including GPU and DRAM power, are measured using the nvidia-smi command in the terminal. The parameters for HDD can be referenced setting in the CPU-...
However, defensive distillation has shown promising results in robustifying image classification deep learning (DL) models against adversarial attacks at the inference level, but they remain vulnerable to data poisoning attacks. This work incorporates a data denoising and reconstruction framework with a ...
The generation and optimization of simulation data for electrical machines remain challenging, largely due to the complexities of magneto-static finite element analysis. Traditional methodologies are not only resource-intensive, but also time-consuming. Deep learning models can be used to shortcut these...
This paper introduces a novel approach to facilitate the construction of surrogate models and selection of informative samples in high-dimensional reliability analysis, through an active learning method based on a deep adversarial autoencoder-based sufficient dimension reduction (AAE-SDR) neural network. ...
using a scale of 10 to study the effects of different weight values on the learning process. The results of training with these five different combinations of weight values are shown in Fig.4. Parallel VAE loss 2 had the highest age prediction accuracy out of all the prediction models ...