Found 10489 publications. Showing page 334 of 420:
Sb-PiPLU: A Novel Parametric Activation Function for Deep Learning
The choice of activation function—particularly non-linear ones—plays a vital role in enhancing the classification performance of deep neural networks. In recent years, a variety of non-linear activation functions have been proposed. However, many of these suffer from drawbacks that limit the effectiveness of deep learning models. Common issues include the dying neuron problem, bias shift, gradient explosion, and vanishing gradients. To address these challenges, we introduce a new activation function: Softsign-based Piecewise Parametric Linear Unit (Sb-PiPLU). This function offers improved non-linear approximation capabilities for neural networks. Its piecewise, parametric design allows for greater adaptability and flexibility, which in turn enhances overall model performance. We evaluated Sb-PiPLU through a series of image classification experiments across various Convolutional Neural Network (CNN) architectures. Additionally, we assessed its memory usage and computational cost, demonstrating that Sb-PiPLU is both stable and efficient in practical applications. Our experimental results show that Sb-PiPLU consistently outperforms conventional activation functions in both classification accuracy and computational efficiency. It achieved higher accuracy on multiple benchmark datasets, including CIFAR-10, CINIC-10, MWD, Brain Tumor, and SVHN, surpassing widely-used functions such as ReLU and Tanh. Due to its flexibility and robustness, Sb-PiPLU is particularly well-suited for complex image classification tasks.
2025
Scaling number concentration measurements from bioaerosol monitors using Hirst-type samplers
The instruments used for routine pollen monitoring are gradually changing from traditional impactors with manual data processing to automated pollen monitors using deterministic and/or machine-learning algorithms for data analysis. This manuscript compares pollen number concentration of Alnus sp., Betula sp., Corylus sp., and Poaceae measured by Hirst-type bioaerosol samplers and the SwisensPoleno automated bioaerosol monitor in Switzerland and Norway. Due to physical particle losses and the classification rate of the algorithms being well below unity, scaling factors had to be applied to the measurements of the SwisensPoleno to match those of the Hirst impactor. These scaling factors depended on the geographic location, i.e. differed significantly between Switzerland and Norway. The importance of adjusting the scaling factors according to the location of the monitoring network and the need for reporting the numerical values of these scaling factors in future scientific publications is emphasized.
2026
2024
2023
2023
Norwegian Scientific Committee for Food and Environment (VKM)
2019
2008
2003
2011
2011
2025
School students working as environmental scientists - Global POP - Dioxins in fish with BDS CALUX. NILU PP
2009
Schools taking part in a research project investigating dioxins in fish. From pole to pole
2016
Screening 2004 - uppföljningsprojekt. Analys av oktaklorstyren, flyktiga metylsiloxaner, vissa fenoler och endosulfan. IVL Rapport, B1745
2007
Screening and assessment of contaminants in human blood samples from Northern Norway, North-West Russia and Siberia. Statlig program for forurensningsovervåking. Rapport 963/2006. TA-2184/2006
2006
2023