Viggo Kann

• PhD
• Professor (Full) at KTH Royal Institute of Technology

Stakeholders and researchers in higher education have long debated the consequences of English-medium instruction (EMI); a key assumption of EMI is that student’s academic learning through English should be at least as good as learning through their first language (usually the national language). This study addressed the following question: “What i...

This work spotlights the experiences from ten years of implementing sustainable development in all educational programs at a technical university. With a focus on the critical issue of involving more academics in the work, experiences are shared through an ethnographic account including focus group interviews. ‘Sustainable development’ has been per...

Teaching assistants (TAs) are often used in computer science (CS) courses to conduct tutorials in smaller groups and to be able to provide students with one-to-one help sessions. TAs can also assist with the assessment, which can be time-consuming in large classes. Previous research does, however, indicate that TAs can be poorly prepared for their...

Teaching assistants (TAs), students who assist the faculty, are widely used in computer science (CS) courses. Previous studies have, however, shown that TAs could be poorly prepared and need training. Particularly, an interview study has shown that one of the areas where the TAs experience uncertainty is when assessing students’ oral presentations...

We present an online API to access a number of Natural Language Processing services developed at KTH. The services work on Swedish text. They include tokenization, part-of-speech tagging, shallow parsing, compound word analysis, word inflection, lemmatization, spelling error detection and correction, grammar checking, and more. The services can be...

Programme Integrating Courses (PICs) aim to tie students, teachers and courses in education programmes closer together. In this study, we investigate three PICs, as part of engineering programmes in computer science and media technology. The purpose of this study was to gain a deeper understanding of how students and mentors experience the PICs wit...

If students have a broad spectrum of study skills, learning will likely be positively affected, since they can adapt the way they learn in different situations. Such study skills can be learned in, for example, learning-to-learn courses. Several studies of such courses have been done over the years, but few of these have been carried out in longitu...

How might the content and outcomes of tertiary education programmes be described and analysed in order to understand how they are structured and function? To address this question we develop a framework for modelling graduate competencies linked to tertiary degree programmes in the computing disciplines. While the focus of our work is computing the...

This is the poster presentation from Working Group 6 "Modeling Competencies for Computing Education" presented at the 23rd Annual Conference on Innovation and Technology in Computer Science Education (ITiCSE'18) in Larnaca, Cyprus. This was a brief summary of the work and preliminary findings of the working group at the conference. A draft report...

This working group contributes to formulating a framework for modeling competencies in the current and future disciplines that comprise computing education. We draw upon the innovative approach taken in the curricular document for information technology (IT2017), curricular competency frameworks, other related documents such as the software enginee...

The aim of this work is to investigate a new method of involving all students in the continued development of an educational program. Using this method, we have obtained a list of well-scrutinized suggestions for improvement that have support among the students, and that we can start to implement. We have also saved a large pool of suggestions that...

The aim of this work is to investigate a new method of involving all students in the continued development of an educational program. Using this method, we have obtained a list of well-scrutinized suggestions for improvement that have support among the students, and that we can start to implement.

How might the content and outcomes of tertiary education programmes be described and analysed in order to understand how they are structured and function? To address this question we develop a framework for modelling graduate competencies linked to tertiary degree programmes in the computing disciplines. While the focus of our work is computing the...

When compared to earlier programming and data structure experiences that our students might have, the perspective changes on computers and programming when introducing theoretical computer science into the picture. Underlying computational models need to be addressed, and mathematical tools employed, to understand the quality criteria of theoretica...

In Sweden, there has since 2013 been a debate in public media, where university professors, mostly from departments of history, have argued that today's students entering university are much less accomplished than earlier students when it comes to basic Swedish language skills. According to the debate, both the spelling and grammar of Swedish stude...

In 2014 the Swedish Language Technology Terminology Group was created, with representatives from different parts of the language technology community, both higher education and research, industry and governmental agencies. In 2016 we have recommended Swedish terms for the 270 language technological concepts in the Bank of Finnish Terminology in Art...

Despite many years of research on Swedish language technology, there is still no well-documented standard for Swedish word processing covering the whole spectrum from low-level tokenization to morphological analysis and disambiguation. SWORD is a new initiative within the SWE-CLARIN consortium aiming to develop documented standards for Swedish word...

If students have a broad spectrum of study skills, learning will likely be positively affected, since they can adapt the way they learn in different situations. Such study skills can be learned in for example learning-to-learn courses. Several studies of such courses have been done over the years, but few of these have been carried out in longitudi...

Study programs are a central component of the higher education system, in particular for professional education. However, for many education programs, especially in engineering, a large part of the beginner students fail to reach graduation. A reason for this could be lack of program coherence. In order to strengthen the program coherence of the co...

Sammanfattning—Programsammanhållande kurser där studenter från årskurs 1-3 gemensamt reflekterar över teman med koppling till deras studier och framtida yrkesliv finns på både KTH och Linköpings universitet (LiU). Syftet med kurserna är främst att skapa en helhet i utbildningen och ge förståelse för vad den leder till, genom att få studenterna att...

Sammanfattning—Vi belyser olikheter och likheter i hur grundutbildningen styrs på fyra svenska tekniska högskolor. Vi jämför hur lärare och examinatorer väljs ut, hur medel fördelas och vilken roll programansvariga (eller motsvarande) har. Abstract—We examine how the undergraduate education is managed in four Swedish engineering schools. We compare...

A program integrating course runs over several years, shows the main thread of the program, and its purpose is to en- able students to become more professional in handling their studies, at the same time as the course has a positive effect on the mentors, other courses and the program itself. This is achieved through regular reection seminars where...

NP-completeness is one of the most central concepts in computer science, and has been extensively applied in many diverse application areas. Despite this, students have problems grasping the concept and, more specifically, applying it to new problems. Independently, we have identified these problems at our universities in different countries and cu...

Skolan för datavetenskap och kommunikation KTH E-post: viggo@nada.kth.se Sammanfattning—På civilingenjörsprogrammet i Datateknik på KTH startades 2010 kursen Programsammanhållande kurs i datateknik, som går över hela de tre första åren av programmet. Kursen fyller minst tio olika funktioner, vilka beskrivs i artikeln. Den är inspirerad av en liknan...

Sammanfattning—Under 2011 genomförs en stor utvärdering av samtliga utbildningsprogram på KTH. Projektet går under namnet Education Assessment Exercise (EAE) och är den största enskilda aktiviteten inom KTH:s systematiska kvalitetsarbete under året [2]. Projektet går genom tre huvudstadier: själv-värdering, extern bedömning och uppföljning. Den för...

Automated assessment systems have been employed in computer science (CS) courses at a number of different universities. Such systems are especially applicable in teaching algorithmic problem solving since they can automatically test if an algorithm has been correctly implemented, i.e., that it performs its specified function on a set of inputs. Bei...

KTH har sedan 2005 arbetat med det automatiserade och webbaserade rättningssystemet Kattis som stöd vid undervisning på programmeringskurser. Detta verktyg möjliggör en ny metodik kring laborationer: testdriven utbildning. Vi har sett stora pedagogiska fördelar med metodiken, och tror att andra högskolor kan dra nytta av våra erfarenheter. Syftet m...

This paper describes an attempt to introduce computer lab exercises on NP-completeness proofs in a class already containing computer lab exercises on algorithms and data structures. In the article we are interested in the answer of the following question: Can the students feel that their understanding of theoretical computer science is improved by...

Evaluation of word space models is usually local in the sense that it only considers words that are deemed very similar by the model. We propose a global evaluation scheme based on clustering of the words. A clustering of high quality in an external evaluation against a semantic resource, such as a dictionary of synonyms, indicates a word space mod...

Proceedings of the 17th Nordic Conference of Computational Linguistics NODALIDA 2009. Editors: Kristiina Jokinen and Eckhard Bick. NEALT Proceedings Series, Vol. 4 (2009), 26-33. © 2009 The editors and contributors. Published by Northern European Association for Language Technology (NEALT) http://omilia.uio.no/nealt . Electronically published at Ta...

This paper presents how we adapted a website search engine for cross language information retrieval, using the Uplug word alignment tool for parallel corpora. We first studied the monolingual search queries posed by the visitors of the website of the Nordic council containing six different languages. In order to compare how well different types of...

The study of the approximability properties of NP-hard optimization problems has recently made great advances mainly due to the results obtained in the field of proof checking. The last important breakthrough proves the APX-completeness of several important optimization problems and thus reconciles 'two distinct views of approximation classes: synt...

This paper deals with analyzing compound words in Swedish using various automatic methods. The paper is written in Swedish, but a similar (somewhat shorter) discussion can be found in "Finding the Correct Interpretation of Swedish Compounds: a Statistical Approach" by Sjobergh and Kann, 2004, and a slightly more thorough discussion can be found in...

Building a large dictionary of syn- onyms for a language is a very te- dious task. Hence there exist very few synonym dictionaries for most languages, and those that exist are generally not freely available due to the amount of work that have been put into them. The Lexin on-line dictionary1 is a very popular web-site for transla- tions of Swedish...

Stockholm viggo@nada.kth.se 26 augusti 2004 Sammanfattning Tjänsten Lexin on-line är oerhört populär: i genomsnitt görs mer än tre uppslag-ningar per sekund, dygnet runt. Denna popularitet är en stor resurs som kan utnytt-jas på följande sätt: Om alla människor som slår upp ett ord erbjuds att svara på en enkel fråga om huruvida två ord är synonyma...

This paper treats compound splitting for Swedish, where compounding is productive and very common. A method for splitting compounds and several methods for choosing the correct interpretation of ambiguous compounds are presented. 99% of all compounds are split, 97% of these are correctly interpreted.

Describe your occupation in a few words", is a question answered by 44 000 Swedish twins. Each respondent was then manually categorized according to two established occupation clas-sification systems. Would a clustering algorithm have produced satisfactory results? Usually, this question cannot be answered. The existing quality measures will tell u...

Grammar errors and context-sensitive spelling errors in texts written by second language learners are hard to detect automatically. We have used three different approaches for grammar checking: manually constructed error detection rules, statistical differences between correct and incorrect texts, and machine learning of specific error types. The...

Introduction In many NLP-applications, the robustness of the internal modules of an application is a prerequisite for the success and usability of the system. The term robustness is a bit unclear and vague, but in NLP, it is often used in the sense robust against noisy, illformed, and partial natural language data. The full spectrum of robustness i...

this paper we focus on tools for grammatical analysis, corpus development and how the corpus must be designed to be useful in machine learning and automatic evaluation

This article describes the construction and performance of Granska – a surface-oriented system for grammar checking of Swedish text. With the use of carefully constructed error detection rules, written in a new structured rule language, the system can detect and suggest corrections for a number of grammatical errors in Swedish texts. In this articl...

This article describes the construction and performance of Granska - a surface-oriented system for grammar checking of Swedish text. With the use of carefully constructed error detection rules, written in a new structured rule language, the system can detect and suggest corrections for a number of grammatical errors in Swedish texts. In this articl...

Four fundamental graph problems, Minimum vertex cover, Maximum independent set, Minimum dominating set and Maximum cut, are shown to be APX-complete even for cubic graphs. Therefore, unless P=NP, these problems do not admit any polynomial time approximation scheme on input graphs of degree bounded by three.

perform an analysis that nds the last word form of compounds, which are common in Swedish. The possible tags of the last word form indicates possible tags (and probability estimation) for an unknown compound word. These two analyses are heuristically combined to get estimations of P (w i jt i ), which enables unknown words to work in the model. Thi...

An efficient implementation of a part-of-speech tagger for Swedish is described. The stochastic tagger uses a well-established Markov model of the language. The tagger tags 92% of unknown words correctly and up to 97% of all words. Several implementation and optimization considerations are discussed. The main contribution of this paper is the thoro...

A method for detecting and correcting spelling errors in Swedish text was presented by Domeij, Hollman and Kann (1994). The objectives were to perform very fast detection and correction of errors and to use a full size word list. Our implementation of the method as a C program is called STAVA. We have further refined this method and implemented ran...

An efficient implementation of a part-of-speech tagger for Swedish is described. The stochastic tagger uses a well-established Markov model of the language. The tagger tags 92% of unknown words correctly and up to 97% of all words. Several implementation and optimization considerations are discussed. The main contribution of this paper is the thoro...

The study of the approximability properties of NP-hard optimization problems has recently made great advances mainly due to the results obtained in the eld of proof checking. The last important breakthrough proves the APX-completeness of several important optimization prob- lems and thus reconciles \two distinct views of approximation classes: synt...

IN PREVIOUS chapters the performance of an approximate algorithm has been analyzed using a worst case point of view. In this chapter this condition is relaxed and the probabilistic analysis of an algorithm is performed by considering either the average case or the behavior in almost all cases.

IN INTRODUCTORY computer programming courses we learn that computers are used to execute algorithms for the solution of problems. Actually, the problems we want to solve by computer may have quite varying characteristics. In general, we are able to express our problem in terms of some relation P⊆ I × S, where I is the set of problem instances and S...

IN THE first chapter we have seen that, due to their inherent complexity, Np-hard optimization problems cannot be efficiently solved in an exact way, unless P = Np. Therefore, if we want to solve an Np-hard optimiza­tion problem by means of an efficient (polynomial-time) algorithm, we have to accept the fact that the algorithm does not always retur...

IN THIS chapter we first present a formal treatment of the complexity concepts introduced in Chap. 1. This treatment will provide the reader with a precise machine-based characterization that will be used in Sect. 6.3 to develop the notion of probabilistically checkable proof (in short, PCP). In Sect. 6.4 we will see how probabilistically checkable...

IN THIS chapter we will give a complete proof of the PCP theorem, that Theorem 6.2. In particular, after a brief presentation of the main ingredients of a Pcp result, we first show that any NP problem admits long (that is, exponential) proofs which can be verified with a constant number of queries. To this aim we introduce the arithmetization of a...

RANDOMIZATION IS one of the most interesting and useful tools in designing efficient algorithms. Randomized algorithms, indeed, have been proposed for many problems arising in different areas: taking into account the scope of this book, however, we will limit ourselves to considering randomized approximation algorithms for Np-hard optimization prob...

IN THE previous two chapters we have seen examples of Npo problems that can be approximated either within a specific constant factor or within any constant factor. We also saw examples of Npo problems for which no approximation algorithm exists (unless P=Np) and examples of Npo problems for which an approximation algorithm but no approxima­tion sch...

IN THE PRECEDING chapters the problem of computing good solutions for Np-hard optimization problems has always been considered in terms of deriving algorithms with a guaranteed behaviour, where such a guarantee refers both to the quality of the returned solution (in terms of either worst case or expected performance ratio) and to the running time (...

WE HAVE seen in the preceding chapters that, even though the decision problems corresponding to most Np-hard optimization problems are polynomial-time Karp-reducible to each other, the optimization problems do not share the same approximability properties. The main reason of this fact is that Karp-reductions not always preserve the measure function...

IN THE preceding chapter we observed that many relevant optimization problems are Np-hard, and that it is unlikely that we will ever be able to find efficient (i.e., polynomial-time) algorithms for their solution. In such cases it is worth looking for algorithms that always return a feasible solution whose measure is not too far from the optimum.

A compendium of NP optimization problems, containing the best approximation results known for each problem, is available on the world wide web at http://www.nada.kth.se/~viggo/problemlist/In this paper we describe the compendium, and specify how the compendium is consultable as well as modifiable on the web. We also give statistics for the use of t...

We investigate the computational complexity of two closely related classes of combinatorial optimization problems for linear systems which arise in various fields such as machine learning, operations research and pattern recognition. In the first class (Min ULR) one wishes, given a possibly infeasible system of linear relations, to find a solution...

A method for detecting and correcting spelling errors in Swedish text was presented by Domeij, Hollman and Kann (1994). The objectives were to perform very fast detection and correction of errors and to use a full size word list. Our implementation of the method as a C program is called Stava. We have further reøned this method and implemented rank...

We investigate how to construct an efficient method for spelling error detection and correction under the prerequisite of using a word list that is encoded and not possible to decode. Our method is probabilistic and the word list is stored as a Bloom filter. In particular we study how to handle compound words and inflections in Swedish. Keywords: s...

We study the following variant of the Max k-Cut problem. Given an input graph G with positively weighted edges and k colors, we wish to compute a subgraph H of G containing “lots” of heavy edges and a color assignment c : V → [k] such that: 1.(a) all edges in H are properly colored and2.(b) a “large fraction” of edges in G n0 ⊮ is properly colored....

A compendium of NP optimization problems, containing the best approximation results known for each problem, is available on the world wide web at http://www.nada.kth.se/˜viggo/problemlist/ In this paper we describe the compendium, and specify how the compendium is consultable as well as modifiable on the web. We also give statistics for the use of...

this paper we obtain new results on the structure of two important computationally-defined classes: the class NPO (that is, the class of optimization problems whose underlying decision problem is in NP), and the class APX (that is, the class of constant-factor approximable NPO problems). In particular, we give the first examples of natural APX-inte...

A compendium of NP optimization problems, containing the best approximation results known for each problem, is now available on the web at http://www.nada.kth.se/viggo/problemlist/. In this paper we describe such a compendium, and specify how the compendium is consultable (and modifiable) on the world wide web.

. We study the Max k-Cut problem and its dual, the Min k-Partition problem. In the Min k-Partition problem, given a graph G = (V; E) and positive edge weights, we want to find an edge set of minimum weight whose removal makes G k-colorable. For the Max k-Cut problem we show that, if P 6= NP, no polynomial time approximation algorithm can achieve a...

Four fundamental graph problems, Minimum vertex cover, Maximum independent set, Minimum dominating set and Maximum cut, are shown to be APX-complete even for cubic graphs. This means that unless P=NP these problems do not admit any polynomial time approximation scheme on input graphs of degree bounded by three.

this paper we shrink the gap for the Max 3-Set Splitting problem and its generalization Max k-Set Splitting from both ends, finding new lower bounds and new upper bounds. In the Max k-Set Splitting problem, the instance consists of subsets of size k of a finite set of elements. The problem is to partition the elements into two parts, such that as m...

We investigate the computational complexity of two classes of combinatorial optimization problems related to linear systems and study the relationship between their approximability properties. In the first class (Min ULR) one wishes, given a possibly infeasible system of linear relations, to find a solution that violates as few relations as possibl...

We study the combinatorial problem which consists, given a system of linear relations, of finding a maximum feasible subsystem, that is a solution satisfying as many relations as possible. The computational complexity of this general problem, named Max FLS, is investigated for the four types of relations =, ⩾, > and ≠. Various constrained versions...

The approximability of several NP maximization problems is investigated and strong lower bounds for the studied problems are proved. For some of the problems the bounds are the best that can be achieved, unless P = NP. For example we investigate the approximability of Max PB 0 – 1 Programming, the problem of finding a binary vector x that satisfies...

Min PB is the class of minimization problems whose objective functions are bounded by a polynomial in the size of the input. We show that there exist several problems that are Min PB-complete with respect to an approximation preserving reduction. These problems are very hard to approximate; in polynomial time they cannot be approximated within n "...

We investigate the computational complexity of the problem which consists, given a system of linear relations, of finding a solution violating as few relations as possible while satisfying all the others. This general combinatorial problem, referred to as Min ULR, is considered for the four basic types of relational operators =, , ? and 6=. We prov...

We prove that maximum H-matching (the problem of determining the maximum number of node-disjoint copies of the fixed graph H contained in a variable graph) is a MAX SNP-hard problem for any graph H that has three or more nodes in some connected component. If H is connected and the degrees of the nodes in H are bounded by a constant the problem is M...

We consider the combinatorial problem MAXFLS which consists, given a system of linear relations, of finding a maximum feasible subsystem, that is a solution satisfying as many relations as possible. The approximability of this general problem is investigated for the three types of relations =, and >. Various constrained versions of MAXFLS where a s...

Min PB is the class of minimization problems whose objective functions are bounded by a polynomial in the size of the input. We show that there exist several problems which are Min PB-complete with respect to an approximation preserving reduction. These problems are very hard to approximate; in polynomial time they cannot be approximated within n f...

Some versions of the maximum common subgraph problem are studied and approximation algorithms are given. The maximum bounded common induced subgraph problem is shown to be Max SNP-hard and the maximum unbounded common induced subgraph problem is shown to be as hard to approximate as the maximum independent set problem. The maximum common induced co...