In recent years, there has been a growing interest in the development of in vitro models to predict cellular behavior within living organisms. Mathematical models, based on differential equations and associated numeri...
详细信息
Through the in-depth analysis of artificial neural network algorithm technology, particle swarm algorithm technology, and image matching algorithm, the article briefly analyzes the theoretical principle of the algorit...
详细信息
The proceedings contain 396 papers. The topics discussed include: development of an IoT enabled smart children tracking and monitoring system using map location services;HMLM: an intelligent artificial intelligence as...
ISBN:
(纸本)9798331543617
The proceedings contain 396 papers. The topics discussed include: development of an IoT enabled smart children tracking and monitoring system using map location services;HMLM: an intelligent artificial intelligence assisted strategy to identify UPI frauds based on hybrid Markov learning methodology;data-driven insights into automotive customer complaints: a machine learning approach to predictive analytics;integrating predictive analytics and deep learning for vehicle safety incident forecasting;analyzing the health status of people in rural areas using machine learning algorithms;machine learning based approach using hand gestures for mouse and video control;experimental analysis of artificial intelligence powered adaptive learning methodology using enhanced deep learning principle;a comprehensive analysis of imageprocessing methods for agricultural product quality control;and convolutional neural network-based multi-fruit classification and quality grading with a Gradio interface.
Skin cancer is a severe health issue. Thus, the major concern of physicians is to investigate a precise clinical diagnosis. At present, some mechanisms are developed in the area of imageprocessing with the help of al...
详细信息
Efficient transfer learning algorithms are key to the success of foundation models on diverse downstream tasks even with limited data. Recent works of Basu et al. (2023) and Kaba et al. (2022) propose group averaging ...
详细信息
ISBN:
(纸本)9781713899921
Efficient transfer learning algorithms are key to the success of foundation models on diverse downstream tasks even with limited data. Recent works of Basu et al. (2023) and Kaba et al. (2022) propose group averaging (equitune) and optimization-based methods, respectively, over features from group-transformed inputs to obtain equivariant outputs from non-equivariant neural networks. While Kaba et al. (2022) are only concerned with training from scratch, we find that equitune performs poorly on equivariant zero-shot tasks despite good finetuning results. We hypothesize that this is because pretrained models provide better quality features for certain transformations than others and simply averaging them is deleterious. Hence, we propose lambda-equitune that averages the features using importance weights, lambda s. These weights are learned directly from the data using a small neural network, leading to excellent zero-shot and finetuned results that outperform equitune. Further, we prove that lambda-equitune is equivariant and a universal approximator of equivariant functions. Additionally, we show that the method of Kaba et al. (2022) used with appropriate loss functions, which we call equizero, also gives excellent zero-shot and finetuned performance. Both equitune and equizero are special cases of lambda-equitune. To show the simplicity and generality of our method, we validate on a wide range of diverse applications and models such as 1) image classification using CLIP, 2) deep Q-learning, 3) fairness in natural language generation (NLG), 4) compositional generalization in languages, and 5) image classification using pretrained CNNs such as Resnet and Alexnet.
The proliferation of surface waste in water bodies poses significant environmental and ecological challenges. Traditional methods of waste detection are often labor-intensive and limited in scope. This paper presents ...
详细信息
ISBN:
(纸本)9798350352900;9798350352894
The proliferation of surface waste in water bodies poses significant environmental and ecological challenges. Traditional methods of waste detection are often labor-intensive and limited in scope. This paper presents a novel approach to surface waste detection using artificial intelligence (AI) and advanced imaging technologies. Leveraging cutting-edge techniques such as deep learning algorithms, high-resolution satellite imagery, and real-time data processing, our system offers an automated solution for identifying and monitoring waste in water bodies. We developed a robust AI model trained on diverse datasets, including satellite and drone-captured images, to detect various types of surface waste with high accuracy. The system integrates real-time processing capabilities to provide timely alerts and actionable insights for environmental management. Evaluation results demonstrate that our approach significantly improves detection accuracy and operational efficiency compared to conventional methods. This research contributes to the advancement of smart environmental monitoring systems and offers a scalable solution for mitigating the impact of surface waste on aquatic ecosystems.
Super-resolution has advanced significantly in the last 20 years, particularly with the application of deep learning methods. One of the most important imageprocessing methods for boosting an image's resolut...
详细信息
As a major breakthrough in artificial intelligence and deep learning, Convolutional Neural Networks have achieved an impressive success in solving many problems in several fields including computer vision and image pr...
详细信息
ISBN:
(数字)9781665496209
ISBN:
(纸本)9781665496209
As a major breakthrough in artificial intelligence and deep learning, Convolutional Neural Networks have achieved an impressive success in solving many problems in several fields including computer vision and imageprocessing. Real-time performance, robustness of algorithms and fast training processes remain open problems in these contexts. In addition object recognition and detection are challenging tasks for resource-constrained embedded systems, commonly used in the industrial sector. To overcome these issues, we propose a dimensionality reduction framework based on Proper Orthogonal Decomposition, a classical model order reduction technique, in order to gain a reduction in the number of hyperparameters of the net. We have applied such framework to SSD300 architecture using PASCAL VOC dataset, demonstrating a reduction of the network dimension and a remarkable speedup in the fine-tuning of the network in a transfer learning context.
Smart homes are to be protected from fire hazards which is a crucial safety concern. Existing ways of detecting fire is time consuming, hence causing maximum injuries and financial loss, so we have come up with an eff...
详细信息
In electronic factors manufacturing, icing product quality through disfigurement discovery is pivotal for maintaining trustability and performance norms. This exploration investigates the operation of deep literacy wa...
详细信息
暂无评论