Distributed, large-scale computing is typically performed using textual general-purpose programming languages. This requires significant programming skills associated with the parallelisation and distribution of computations. In this paper, we present a visual (graphical) programming language called the Computation Application Language (CAL) to raise abstraction in distributed computing. CAL programs define computation workflows by visualising data flowing between computation units. The goal is to reduce the amount of traditional code needed and thus facilitate development even by non-professional programmers. The language follows the low-code paradigm, i.e. its implementation (the editor and the runtime system) is available online. We formalise the language by defining its syntax using a metamodel and specifying its semantics using a two-step approach. We define a translation of CAL into an intermediate language which is then defined using an operational approach. This formalisation was used to develop a programming and execution environment. The environment orchestrates computations by interpreting the intermediate language and managing the instantiation of computation modules using data tokens. We also present an explanatory case-study example that shows a practical application of the language.

Background: So far there has been little evidence that implementation of the health information technologies (HIT) is leading to health care cost savings. One of the reasons for this lack of impact by the HIT likely lies in the complexity of the business process ownership in the hospitals. The goal of our research is to develop a business model-based method for hospital use which would allow doctors to retrieve directly the ad-hoc information from various hospital databases. Methods: We have developed a special domain-specific process modelling language called the MedMod. Formally, we define the MedMod language as a profile on UML Class diagrams, but we also demonstrate it on examples, where we explain the semantics of all its elements informally. Moreover, we have developed the Process Query Language (PQL) that is based on MedMod process definition language. The purpose of PQL is to allow a doctor querying (filtering) runtime data of hospital’s processes described using MedMod. Results: The MedMod language tries to overcome deficiencies in existing process modeling languages, allowing to specify the loosely-defined sequence of the steps to be performed in the clinical process. The main advantages of PQL are in two main areas – usability and efficiency. They are: 1) the view on data through “glasses” of familiar process, 2) the simple and easy-to-perceive means of setting filtering conditions require no more expertise than using spreadsheet applications, 3) the dynamic response to each step in construction of the complete query that shortens the learning curve greatly and reduces the error rate, and 4) the selected means of filtering and data retrieving allows to execute queries in O(n) time regarding the size of the dataset. Conclusions: We are about to continue developing this project with three further steps. First, we are planning to develop user-friendly graphical editors for the MedMod process modeling and query languages. The second step is to do evaluation of usability the proposed language and tool involving the physicians from several hospitals in Latvia and working with real data from these hospitals. Our third step is to develop an efficient implementation of the query language.

Convolutional neural networks have become the main tools for processing two-dimensional data. They work well for images, yet convolutions have a limited receptive field that prevents its applications to more complex 2D tasks. We propose a new neural model, called Matrix Shuffle-Exchange network, that can efficiently exploit long-range dependencies in 2D data and has comparable speed to a convolutional neural network. It is derived from Neural Shuffle-Exchange network and has $\mathcal{O}( \log{n})$ layers and $\mathcal{O}( n^2 \log{n})$ total time and space complexity for processing a $n \times n$ data matrix. We show that the Matrix Shuffle-Exchange network is well-suited for algorithmic and logical reasoning tasks on matrices and dense graphs, exceeding convolutional and graph neural network baselines. Its distinct advantage is the capability of retaining full long-range dependency modelling when generalizing to larger instances - much larger than could be processed with models equipped with a dense attention mechanism.