BIBSYS: Open Journals Systems
Not a member yet
393 research outputs found
Sort by
Bias mitigation with AIF360: A comparative study
Get PDFThe use of artificial intelligence for decision making raises concerns about the societal impact of such systems. Traditionally, the product of a human decision-maker are governed by laws and human values. Decision-making is now being guided - or in some cases, replaced by machine learning classification which may reinforce and introduce bias. Algorithmic bias mitigation is explored as an approach to avoid this, however it does come at a cost: efficiency and accuracy. We conduct an empirical analysis of two off-the-shelf bias mitigation techniques from the AIF360 toolkit on a binary classification task. Our preliminary results indicate that bias mitigation is a feasible approach to ensuring group fairness
Specifying Software Languages: Grammars, Projectional Editors, and Unconventional Approaches
Get PDFWe discuss several approaches for defining software languages, together with Integrated Development Environments for them. Theoretical foundation is grammar-based models: they can be used where proven correctness of specifications is required. From a practical point of view, we discuss how language specification can be made more accessible by focusing on language workbenches and projectional editing, and discuss how it can be formalized. We also give a brief overview of unconventional ideas to language definition, and outline three open problems connected to the approaches we discuss
Editorial for the IARTEM e-journal, July 2020
Get PDFEditorial for the IARTEM e-journal 2020, Volume 12 No. 1
 
Some Faster Algorithms for Finding Large Prime Gaps
Get PDFThis paper investigates the problem of finding large prime gaps (the difference between two consecutive prime numbers, pi+1 – pi) and on the development of a small, efficient program for generating such large prime gaps for a single computer, a laptop or a workstation. In Wikipedia [1], one can find a table of all known record prime gaps less than 264, the record is a 20 decimal digit number. We wanted to go beyond 64 bit numbers and demonstrate algorithms that do not needed a huge number of computers in a grid to produce useful results.
After some preliminary tests, we found that the Sieve of Eratosthenes, SE, from the year 250 BC was the fastest for finding prime numbers and it could also be made space efficient. Each odd number is represented by one bit and when storing 8 odd numbers in a single byte (representing 16 consecutive numbers ignoring the even numbers), we found that we should not make one long SE table, but instead divide the SE table into segments (called SE segments), each of length 108 or 109 and dynamically generate the necessary SE segments as to find prime numbers. First, we made a basic segment of all prime numbers < 108 (in less than a second). We also relied heavily on the old observation [2] that when using SE to find all prime numbers ?????, we cross out all numbers using the prime numbers ???? ? ?????, and that the first number crossed off when crossing out for prime number p is p2. When we want to find prime gaps, we first create one or more consecutive SE in that range, say starting on 274 and ending with the value M – initially these big segments are crossed out by our first basic set of primes < 108 , To find all prime number in these big segments, we next need the rest of prime numbers ???? ? ????? . These can be all be constructed by using our first set of prime numbers to generate segments of consecutive SE from 108. The primes in these segments are used to cross out in the big SE segment and can then be discarded (each prime used only once). Our most significant algorithm was to find a simple formula for using primes from a range 3 – 236 to cross out the non-primes in any SE segment without crossing out in all the numbers between 236 and 272. This leads to an exponential saving in both space and execution time. In addition to this, we created a small package Int3 to represent numbers > 264 by storing 8 decimal values in each of 3 integer variables together with the necessary mathematical operations. The Int3 package can handle numbers up to 24 decimal digits and is significantly faster than the BigInteger package in the Java library. We also created a faster algorithm for finding all record prime gaps.
The results presented in this paper are some tables of prime gaps for primes significantly larger than 264 and data supporting an observation that big prime gaps in these segments are much more frequent than the ones we find in the Wikipedia table where the search starts at prime number 3. Our combined set of algorithms is also sufficiently fast to test every entry in the Wikipedia table in less than 5 minutes. We conclude by reflecting on the use of brute force (more computers) versus smarter algorithms
Analyse av studenters eksamenskode: typiske feil og muligheter for autoretting
Get PDFSelv om programkode egner seg for automatisk prosessering, har den typisk vært vurdert manuelt, også etter innføring av digital eksamen. Vanlige verktøy for digital eksamen har ikke tilbudt oppgavetyper hvor studentene kan teste koden underveis, ei heller at sensor kjører koden under retting. Slik funksjonalitet vil bli tilgjengelig i nær framtid, og det er interessant å se hvilke implikasjoner dette kan ha for eksamen. Denne artikkelen analyserer studentkode på noen forholdsvis enkle oppgaver i Python-programmering fra en eksamen i høst 2019. Få studenter hadde kode som ville ha kjørt feilfritt, men vesentlig flere hadde kode som de kunne ha klart å rette til feilfri kjøring i løpet av kort tid gitt testmulighet. Analysen avdekker også typiske feil som gikk igjen i mange studenters kode, og diskuterer hvordan en eksamen med helt eller delvis automatisk retting av koden vil slå ut
The Communicative Competence Elements in the Foreign Language Textbooks: A Descriptive Case Study on Turkish and English Textbooks
Get PDFThe textbooks used in teaching foreign language are considered as important tools since the more quality a textbooks is, the better learning or teaching occurs. Therefore, the studies conducted on analyzing and evaluating the textbooks have contributed to the field of developing textbooks. Thus our aim is to analyze and compare the use of communicative competence components, namely pragmatic competence, discourse competence and strategic competence in the speaking activities of two textbooks used in teaching English and Turkish as a foreign language. We adopted a descriptive case study design within the qualitative research framework to serve to the purpose of the study, and analyzed the textbooks through the combination of impressionistic, checklist, and in-depth-internal evaluation methods. The comparative analysis of the textbooks revealed that although the pragmatic competence component is covered pleasingly, we found the discourse competence component is covered more effectually in the textbook used in teaching English as a foreign language. Additionally we found that neither of the books are particularly weak in covering the strategic competence aspects in the speaking activities. Based on our findings, we highly suggest to the textbook writers and publishers to pay more attention to such weak points in order to simulate the natural language use and to increase the communication functions of the textbooks used in foreign language teaching
Evaluating multi-core graph algorithm frameworks
Get PDFMulti-core and GPU-based systems offer unprecedented computational power. They are, however, challenging to utilize effectively, especially when processing irregular data such as graphs. Graphs are of great interest, as they are now used to model geographic-, social- andneural networks. Several interesting programming frameworks for graph processing have therefore been developed these past few years.
In this work, we highlight the strengths and weaknesses of the Galois, GraphBLAST, Gunrock and Ligra graph frameworks through benchmarking their single source shortest path (SSSP) implementations using the SuiteSparse Matrix Collection. Tests were done on an Nvidia DGX2 system, except for Ligra, which only provides a multi-core framework. D-IrGL, built on Galois, also provided a multi-GPU option for SSSP. We also look at program size, documentation and overall ease of use.
High performance generally comes at the price of high complexity. D-IrGL shows its strength on the very largest graphs, where it achieved the best run-time, while Gunrock processed most other large sets the fastest. However, GraphBLAST, with a relatively low-complexity interface, achieves the greatest median throughput across all our test cases. This despite that its SSSP implementation size is only 1/10th of Gunrock, which for our tests has the highest peak throughput and the fastest run-time in most cases. Ligra had less computational resources available, and consequently performed worse in most cases, but it is also a very compact and easy to use framework. Futher analyses and some suggestions for future work are also included
Deep Active Learning for Autonomous Perception
Get PDFTraditional supervised learning requires significant amounts of labeled training data to achieve satisfactory results. As autonomous perception systems collect continuous data, the labeling process becomes expensive and time-consuming. Active learning is a specialized semi-supervised learning strategy that allows a machine learning model to achieve high performance using less training data, thereby minimizing the cost of manual annotation. We explore active learning for autonomous vehicles, and propose a novel deep active learning framework for object detection and instance segmentation. We review prominent active learning approaches, study their performances in the aforementioned computer vision tasks, and perform several experiments using state-of-the-art R-CNN-based models for datasets in the self-driving domain. Our empirical experiments on a number of datasets reflect that active learning reduces the amount of training data required. We observe that early exploration with instance-rich training sets leads to good performance, and that false positives can have a negative impact if not dealt with appropriately. Furthermore, we perform a qualitative evaluation using autonomous driving data collected from Trondheim, illustrating that active learning can help in selecting more informative images to annotate
Teaching AI Ethics: Observations and Challenges
Get PDFThis report summarises the experience in teaching Artificial Intelligence (AI) Ethics as an elective masters level course at the University of Bergen. The goal of the summary is twofold: 1) to draw lessons for teaching this in-high demand very new discipline; 2) to serve as a basis in developing a bachelor level AI Ethics course for students of artificial intelligence. AI Ethics as a topic is particularly challenging to teach as the discipline itself is very new and no textbooks have been established. The added challenge is introducing methodologies and skills from humanity- and social sciences to students of computational and information sciences
Incorporating societal topics in software engineering education: A case study of a customer-driven course
Get PDFContext: This full research paper presents a review of the project descriptions from a projectbased course designed around close collaboration with students and external customers. Our master course is based on four decades of software engineering experience and teaching at Norwegian University of Science and Technology. In the scope of this work, we analyzed 45 customers’ project proposals over the past three years.
Objectives: More precisely, we looked into 1) how many societal topics are present in project customers’ descriptions? 2) Which sustainable development goals if any, are addressed from the customers? 3) How do the trends of societal topics addressing SDGs in customer projects change over time? 4) Which categories of individuals do project descriptions target?
Methods: We conducted a deductive thematic analysis utilizing open coding of the customers’ project descriptions. Results: We found that most project descriptions provided by the customers had a technical focus with a moderate portion of projects addressing societal topics for specific target groups.
Contribution: The study’s overall outcomes contribute to the course’s future improvement and informs customers about the prospective socially relevant project proposals