Authored by Lucy Frith, Marie-Clare Balaam, Soo Downe July 2020 sees the launch of the PLOS Special Collection Understanding childbirth as a…
By Ruxandra Stoean, Associate Professor of Computer Science, University of Craiova, Leonardo Franco, Associate Professor of Computer Science, University of Malaga, Miguel Atencia, Associate Professor of Applied Mathematics, University of Malaga, Gonzalo Joya, Professor of Electronics Technology, University of Malaga.
The PLOS IWANN Special Collection 2019: Deep Learning Models in Healthcare and Biomedicine was built upon papers previously selected from the special session dedicated to health and biomedical applications of deep learning at the 15th International Work-Conference on Artificial Neural Networks (IWANN 2019).
Medicine seems like the perfect candidate for deep learning exploration given the huge amount of complex data that can be obtained from the human body. As such, there is a continuous interest in tailoring deep networks to discover the most significant variables, the complex interactions among them, as well as models within the field. The session welcomed recent contributions in the areas connected to the design and application of deep learning techniques for healthcare tasks and biomedical studies. A primary target of this call was to tackle the issue that often the data needs to be reshaped first for a better interplay with the deep neural network models.
The articles in the Special Collection address a wide variety of deep architectures, such as convolutional, long short-term memory, and recurrent neural networks. Additional refinements were performed on the deep architectures: hyper-parameter tuning by genetic algorithms, vector-to-image encoding, transfer learning, data pre-processing through clustering.
Beyond typical deep learning medical applications that target image data, those in the current collection were focused on signal and gene-expression samples. On the one hand, electroencephalography time series were modelled for Alzheimer detection and motor imagery classification, and electrooculography signals were analyzed for supporting diagnosis of spinocerebellar ataxia type 2. On the other hand, gene-expression data of tens of distinct tumour types were studied for cancer survival prediction.
You are invited to check out the papers in the IWANN 2019 Special Collection: https://collections.plos.org/iwann-2019
High level research on biomedical applications of deep learning is much needed in these challenging times. Thus we are looking forward to welcoming your submission to the next IWANN conference in 2021: http://iwann.uma.es/
Featured Image Caption: The 3.6m deep Ochiul Beiului Lake, Romania.
Featured Image Credit: Ruxandra Stoean, Associate Professor of Computer Science, University of Craiova.