Artificial intelligence is currently a hot topic in medicine. However, medical data is often sparse and hard to obtain due to legal restrictions and lack of medical personnel for the cumbersome and tedious process to manually label training data. These constraints make it difficult to develop systems for automatic analysis, like detecting disease or other lesions. In this respect, this article presents HyperKvasir, the largest image and video dataset of the gastrointestinal tract available today. The data is collected during real gastro- and colonoscopy examinations at Bærum Hospital in Norway and partly labeled by experienced gastrointestinal endoscopists. The dataset contains 110,079 images and 374 videos, and represents anatomical landmarks as well as pathological and normal findings. The total number of images and video frames together is around 1 million. Initial experiments demonstrate the potential benefits of artificial intelligence-based computer-assisted diagnosis systems. The HyperKvasir dataset can play a valuable role in developing better algorithms and computer-assisted examination systems not only for gastro- and colonoscopy, but also for other fields in medicine.

Modern workloads often exceed the processing and I/O capabilities provided by resource virtualization, requiring direct access to the physical hardware in order to reduce latency and computing overhead. For computers interconnected in a cluster, access to remote hardware resources often requires facilitation both in hardware and specialized drivers with virtualization support. This limits the availability of resources to specific devices and drivers that are supported by the virtualization technology being used, as well as what the interconnection technology supports. For PCI Express (PCIe) clusters, we have previously proposed Device Lending as a solution for enabling direct low latency access to remote devices. The method has extremely low computing overhead and does not require any application- or device-specific distribution mechanisms. Any PCIe device, such as network cards disks, and GPUs, can easily be shared among the connected hosts. In this work, we have extended our solution with support for a virtual machine (VM) hypervisor. Physical remote devices can be “passed through” to VM guests, enabling direct access to physical resources while still retaining the flexibility of virtualization. Additionally, we have also implemented multi-device support, enabling shortest-path peer-to-peer transfers between remote devices residing in different hosts. Our experimental results prove that multiple remote devices can be used, achieving bandwidth and latency close to native PCIe, and without requiring any additional support in device drivers. I/O intensive workloads run seamlessly using both local and remote resources. With our added VM and multi-device support, Device Lending offers highly customizable configurations of remote devices that can be dynamically reassigned and shared to optimize resource utilization, thus enabling a flexible composable I/O infrastructure for VMs as well as bare-metal machines.

This paper details the approach to the MediaEval 2018 Medico Multimedia Task made by the Rune team. The decided upon approach uses a work-in-progress hyperparameter optimization system called Saga. Saga is a system for creating the best hyperparameter finding in Keras, a popular machine learning framework, using Bayesian optimization and transfer learning. In addition to optimizing the Keras classifier configuration, we try manipulating the dataset by adding extra images in a class lacking in images and splitting a commonly misclassified class into two classes.

The field of medicine has a history of adopting new technology. Video equipment and sensors are used to visualize areas of interest in the human allowing for doctors to make diagnoses based on imagery observations. However, the detection rate of the doctors towards diseases and abnormalities is heavily dependent on the experience and state of mind of the doctor doing the examination. Computer-aided detection systems are systems designed to aid the doctor in improving the detection rate, and they are using or experimenting with machine learning. Deep convolutional neural networks, a type of machine learning, are shown to be highly efficient at image detection, classification, and analysis. However, these networks require large datasets to train properly. Transfer learning is a training technique where we use a pre-trained machine learning model and transfer some of the attained knowledge from other application domains over to a new model. This way, we can use small datasets and train a model in much shorter time. In this respect, transfer learning works fine but has many configurations called hyperparameters which are often not optimized. Our work aims to address the lack of automatic hyperparameter optimization for transfer learning by experiments utilizing a known hyperparameter optimization method and creating a system for running those experiments. First, we decided to focus on the field of gastroenterology by utilizing two publicly available datasets showing images from the gastrointestinal tract. Next, we used a specific transfer learning method and chose hyperparameters suitable for automatic optimization. The optimization method we chose was Bayesian optimization because of its reputation for being one of the best methods for hyperparameter optimization. However, Bayesian optimization has hyperparameters of its own, and there are also different versions of Bayesian optimization. We chose to limit the thesis, so we use standard Bayesian optimization with standard parameters. We created a system for running automatic experiments of three different hyperparameter optimization strategies. With the system, we ran a set of experiments for each dataset. Between the strategies, one was successful in achieving a high validation accuracy, while the others were considered failures. Compared to baselines, our best models was around 10% better. With these experiments, we demonstrated that automatic hyperparameter optimization is an effective strategy for increasing performance in transfer learning and that the best hyperparameters are nontrivial to select manually.