Norsk IKT-konferanse for forskning og utdanning

Students’ mental models of references in Python

Kristin Marie Rørnes — 2019-11-14

This paper reports on a study exploring students’ understanding of references and reference assignments. Students in an introductory programming course in Python were interviewed with respect to what happens during execution of reference assignment statements and function calls involving references. Previous research on Java has
identified two types of mental models related to reference assignment, which in this paper is referred to as the “Copy value model” and the “Copy reference model”.
An important result in this paper is that each of these models can be divided into two sub-models, giving a total of four different models where only one of them is valid. In addition, we have identified three types of mental models related to references in function calls. This gives valuable insight into students’ thinking, which can then be addressed by teachers both in class and in formative assessments. Furthermore, students in two introductory courses were asked to participate in a survey with multiple choice questions asking the students to identify the correct results of executing code examples involving references. The patterns of answers were analyzed based on the mental models identified in interviews. It was found that the identified mental models explained the most common patterns in the student responses.

Mandatory coursework in higher Norwegian IT education

Per Lauvås jr — 2019-11-14

Obligatory exercise, mandatory activity and work requirement are all examples of terms describing the same phenomenon in higher Norwegian education: Something a student needs to pass in order to get access to an exam. In this paper we call them mandatory coursework in alignment with relevant existing research. Some argue that mandatory coursework assignments can, and should, be eliminated. Before we can discuss this within Norwegian IT education, we need to know to what extent mandatory coursework is in use. A course description should describe any mandatory coursework within a course. In this paper we present extracted data from course descriptions from 12 institutions delivering IT education in Norway. The data tells us the frequency in which mandatory coursework is in use, the different types used, how many there are, in what stages within a study programme they are most commonly in use and the variation between the 12 institutions. The results tell us that mandatory coursework to a large extent is in use in Norwegian IT education, although there are significant variations among the different institutions. The most common coursework are labs, assignments and submissions, but participation is also quite common. Mandatory coursework is in use in both bachelor and master programmes with year one in a study as the most coursework intensive.

VisAST: Generic AST Visualiser for Software Language Education

Ragnhild Aalvik — 2019-11-18

Structural concepts such as abstract syntax trees (ASTs) are often best explained through visual representations. Students may, however, struggle with connecting such visual representations with the corresponding program text. To bridge this gap, we developed visAST, a tool for easily visualising ASTs of small languages written in Haskell. To assess the benets and usability of visAST we conducted a user study in the context of students implementing interpreters. Students reported liking visAST and it being benecial for learning. The experiment's results were not conclusive, but hint at visAST use improving students' performance.

Investigating students’ journey through a computer science program using exam data: three new approaches

Madeleine Lorås — 2019-11-18

A computing student will over the first three years of their studies complete approximately 20 exams and even more attempts due to failures and retakes. Details about all exam attempts are stored in a national database called Common Student System (Felles studentsystem, FS). Although access to FS, in general, is restricted, anonymized data about exam attempts can be provided for research and is a potential goldmine of data that, if used right might be a useful tool for educators and teachers. In this study, we explore this data in an attempt to conceptualize three new approaches to assessing student performance. Firstly, we relate students' final grade point average (GPA) to their performance in all courses in the first two years. Additionally, we propose a new indicator of student performance called "struggle factor," which is calculated using the number of exam attempts. Lastly, we investigate how students perform in different course subjects and types. Both the proposed use of FS data and the new approaches to performance indicators are relevant for educators wanting to understand the educational design of a study program and the students’ journey.

Den tause kunnskapen i IT-studia

Hans Georg Schaathun — 2019-11-18

Michael Polanyi er kjend for omgrepet taus kunnskap. Han viste at me ofte veit meir enn me kan fortelja. Dette ser me tydleg i profesjonsstudia. Studentane treng ein praktisk kompetanse som førelesarane slit med å setja ord på. Denne artikkelen oppsummerer litt av det som har vore skrive om kva taus kunnskap er, kvifor han er taus og korleis han vert formidla. Me argumenterer for at det er den tause kunnskapen som skil menneske frå maskin og dermed er det som studentane våre treng i ein kvar jobb som ikkje snart vert robotisert.

Pass/Fail Grading and Educational Practices in Computer Science

Torstein J. F. Strømme — 2019-12-03

Binary (pass/fail) grading have been shown to have benefits with respect to mental health and collaboration, and is argued to promote a deep approach to learning. However, diverging results with respect to academic achievement suggests that the full benefits of binary grading are contingent on underlying factors, such as how the teaching and learning activities in the course are designed. We here present experiences and student feedback for an intermediate level course in computer science that is graded using pass/fail, and which is highly successful both in terms of of student enjoyment and academic achievement. Survey results also indicate that students apply a deeper learning approach towards the course than average. Drawing on examples and findings from this course, we argue that the following three practices makes a binary graded course in computer science successful: a) a sufficiently high bar for passing, b) clear course requirements, and c) the use of formative assessment.

Forord til NIK 2019 og UDIT 2019

Arne Lakså — 2019-11-18

Blending functions based on trigonometric and polynomial approximations of the Fabius function

Hans Olofsen — 2019-11-14

Most simple blending functions are polynomials, while more advanced blending functions are, for example, rational or expo-rational fractions. The Fabius function has the required properties of a blending function, but is a nowhere analytic function and cannot be calculated exactly everywhere on the required domain. We present a new set of trigonometric and polynomial blending functions with the shape and other properties similar to the Fabius function. They consist of combinations of trigonometric polynomials and piecewise polynomials. The main advanced of these are that they are easy to implement, have low processing costs and have simple derivatives. This makes them very suitable for the calculation of splines. Due to the selfdifferential property of the Fabius function, scaled versions of these functions can even be used to approximate their own derivatives.

Escape Local Minima with Improved Particle Swarm Optimization Algorithm

Kuruge Darshana Abeyrathna — 2019-11-14

Particle Swarm Optimization (PSO) is a powerful meta-heuristic technique which has been maneuvered to solve numerous complex optimization problems. However, due to its characteristics, there is a possibility to trap all particles in a local minimum in the solution space and then they cannot find the way out from the trap on their own. Therefore, we modify the traditional PSO algorithm by adding an extra step so that it helps PSO to find a better solution than the local minimum that they undesirably found. We perturb all the particles by adjusting parameter values in the traditional algorithm when there is no improvement of the objective value over the training iterations, assuming that particles have stuck in a local minimum. In this research, we mainly focus on adjusting the learning factors. However, the parameter values have to be used in an effective way to perturb the particles. The behavior of the proposed modification and its parameter adjustments are studied using a function which has a large number of local minima - Schwefel’s function. Results show that 2 out of 3 PSO attempts trap in local minimum and slight changes on learning factors do not help them to get out from the traps. However, perturbances made with large learning factors can find better solutions than the local minima that they stuck in and help to find the global minimum eventually.

Cloud-based Implementation and Validation of a Predictive Fire Risk Indication Model

Lars Kristensen — 2019-11-14

The high representation of wooden houses in Norwegian cities combined with periods of dry and cold climate during the winter time often results in a high risk of severe fires. This makes it important for public authorities and fire departments to have an accurate estimate of the current fire risk in order to take proper precautions. We report on the implementation of a predictive mathematical model based on first order principles which exploits cloud-provided measurements from weather stations and weather forecasts from the Norwegian Meteorological Institute to predict the current and future fire risk at a given geographical location. We have experimentally validated the model during the winter 2018-2019 at selected geographical locations, and by considering weather data from the time of several historical fires. Our results show that our cloud and web-based implementation is both time and storage efficient, and capable of being able to accurately predict the fire risk measured in terms of the estimated time to ashover. The paper demonstrates that our methodology in the near future may become a valuable risk predicting tool for Norwegian fire brigades.

Autonomous Vehicle Control: End-to-end Learning in Simulated Environments

Hege Haavaldsen — 2019-11-14

This paper examines end-to-end learning for autonomous vehicles in diverse, simulated environments containing other vehicles, traffic lights, and traffic signs; in weather conditions ranging from sunny to heavy rain. The paper proposes an architecture combing a traditional Convolutional Neural Network with a recurrent layer to facilitate the learning of both spatial and temporal relationships. Furthermore, the paper suggests a model that supports navigational input from the user to facilitate the use of a global route planner to achieve a more comprehensive system. The paper also explores some of the uncertainties regarding the implementation of end-to-end systems. Specifically, how a system’s overall performance is affected by the size of the training dataset, the allowed prediction frequency, and the number of hidden states in the system’s recurrent module. The proposed system is trained using expert driving data captured in various simulated settings and evaluated by its real-time driving performance in unseen simulated environments. The results of the paper indicate that end-to-end systems can operate autonomously in simulated environments, in a range of different weather conditions. Additionally, it was found that using ten hidden states for the system’s recurrent module was optimal. The results further show that the system was sensitive to small reductions in dataset size and that a prediction frequency of 15 Hz was required for the system to perform at its full potential.

An automatic image-based system for detecting wild and stocked fish

Espen Myrum — 2019-11-14

Fish stocking is the method of raising fish in a hatchery and releasing them into a river or lake to sustain or increase an existing population or to create a population. This has been practised in many countries, including Norway. Before the fish are released, the adipose fin is commonly removed in order to identify that it is a stocked fish. Cameras have been mounted in several Norwegian rivers in order to monitor fish populations. Classification of fish from these cameras is today a manual task carried out by people. In this paper we propose an automatic classification method to separate wild fish from stocked fish using machine learning. Experiments on an image set of trouts (Salmo Trutta) show a very high accuracy of the proposed method.

Evaluating Population Based Training on Small Datasets

Frode Tennebø — 2019-11-14

Recently, there has been an increased interest in using artificial neural networks in the severely resource-constrained devices found in Internet-of-Things networks, in order to perform actions learned from the raw sensor data gathered by these devices. Unfortunately, training neural networks to achieve optimal prediction accuracy requires tuning multiple hyper-parameters, a process which has traditionally taken many times the computation time of a single training run of the neural network. In this paper, we empirically evaluate the Population Based Training algorithm, a method which simultaneously both trains and tunes a neural network, on datasets of similar size to what we might encounter in an IoT scenario. We determine that the population based training algorithm achieves prediction accuracy comparable to a traditional grid or random search on small datasets, and achieves state-of-the-art results for the Biodeg dataset.

Evaluating accessibility testing in automated software build processes

Aleksander Bai — 2019-11-14

Today, most software projects utilize an automated build process with unit tests, code quality checks, end-to-end integration tests, and more. To make software usable by as many people as possible, regardless of capabilities, accessibility testing must be integrated into the build process. The goal is highly accessible software where issues are found early in the development lifecycle. In this work, we have investigated the most common accessibility testing tools suitable for integration in an automated build process and split the tools into two categories based on their rulesets. The rulesets are evaluated against a well known demonstration site for calculation of the rulesets' precision, recall, and $F_1$ scores. Finally, we discuss the implications of their scores and how accessibility testing tools can be integrated into an automated build process.

Exploring future C++ features within a geometric modeling context

Jostein Bratlie — 2019-11-14

The development of the C++ programming language and its standard library has undergone a renaissance since the introduction of the C++11 standard, making the language more relevant than ever, through modern features, simplification and expansion of the standard library. Comparing past and future feature sets (C++17, C++20, ...) similar to comparing different programming languages. In this article we look at how new and upcoming features of the language can be utilized to ease the development of domain specific application areas through features such as generic programming, compile-time polymorphism and type-safety. We provide representative examples by application to differential geometry by modeling hierarchical structure for parametric object evaluation.

Geological Multi-scenario Reasoning

Crystal Chang Din — 2019-11-14

In the oil and gas industry, during exploration prospect assessment, explorationists rely on ad hoc manual work practices and tools for developing and communicating multiple hypothetical geological scenarios of the prospect. This leaves them with little efficient means to make the fullest use of state of the art digital technologies to communicate and systematically compare and assess different hypothetical geological scenarios before deciding which scenario to pursue. In this paper, we present a formal framework for geological multi-scenario reasoning, a novel tool-based method for geologically oriented subsurface evaluation. The methodology applies formal methods and logic-based techniques to subsurface evaluation and expresses interpretive uncertainty as discrete scenarios with branches of potential alternatives. This framework consists of (i) a proto-scenario generator that takes user observations and geological evidence as input and generates semantically valid initial states based on formalized geological knowledge in first-order logic (ii) geological processes formalized as a rewrite theory that are executable in Maude. By applying geological rewrite rules onto the proto-scenarios, we are able to assist explorationists with multi-scenario generation and reasoning beyond human capacity.