Manufacturing plants are highly dependent on machines and involve a significant number of equipment to produce a finished product. Industry 4.0 helps structure the processes involved in such setups and enables the fun...
详细信息
Zero-shot quantization aims to learn a quantized model from a pre-trained full-precision model with no access to original real training data. The common idea in zero-shot quantization approaches is to generate synthet...
详细信息
Zero-shot quantization aims to learn a quantized model from a pre-trained full-precision model with no access to original real training data. The common idea in zero-shot quantization approaches is to generate synthetic data for quantizing the full-precision model. While it is well-known that deep neural networks with low sharpness have better generalization ability, none of the previous zero-shot quantization works considers the sharpness of the quantized model as a criterion for generating training data. This paper introduces a novel methodology that takes into account quantized model sharpness in synthetic data generation to enhance generalization. Specifically, we first demonstrate that sharpness minimization can be attained by maximizing gradient matching between the reconstruction loss gradients computed on synthetic and real validation data, under certain assumptions. We then circumvent the problem of the gradient matching without real validation set by approximating it with the gradient matching between each generated sample and its neighbors. Experimental evaluations on CIFAR-100 and ImageNet datasets demonstrate the superiority of the proposed method over the state-of-the-art techniques in low-bit quantization settings. Copyright 2024 by the author(s)
A crucial problem in natural language processing is language identification, which has applications in speech recognition, translation services, and multilingual content. The five main Indian languages that are the su...
详细信息
This research bridges traditional Kazakh heritage with modern science by developing a machine learning model to classify Kazakh clans using Y-chromosome data. This paper introduces an innovative approach for assigning...
详细信息
In order to promote the evaluation performance of deep learning infrared automatic target recognition (ATR) algorithms in the complex environment of air-to-air missile research, we proposed an analytic hierarchy proce...
详细信息
Since increasing competition and decreasing resources in the globalizing world, companies have focused on activities that add value to the product and supply chain management has become substantial. On the other hand,...
详细信息
The significant increase in the number of patients with ocular diseases has made medical institutions face great challenges. At present, computer aided diagnosis technology is urgently needed to assist doctors in clin...
详细信息
The recognition of intrusion attempts is the fundamental region of network security, with the objective of identifying the impact of these actions on the distinctive variations of the captured traffic. Innovation and ...
详细信息
Conformal prediction is a powerful tool for uncertainty quantification, but its application to time-series data is constrained by the violation of the exchangeability assumption. Current solutions for time-series pred...
In video analysis, existing approaches achieve excellent performance under preset conditions, but the scarcity of training videos limits deep learning development. To address this, this paper introduces two external r...
详细信息
暂无评论