his.sePublications
Change search
Refine search result
12 1 - 50 of 74
CiteExportLink to result list
Permanent link
Cite
Citation style
  • apa
  • harvard1
  • ieee
  • modern-language-association-8th-edition
  • vancouver
  • Other style
More styles
Language
  • de-DE
  • en-GB
  • en-US
  • fi-FI
  • nn-NO
  • nn-NB
  • sv-SE
  • Other locale
More languages
Output format
  • html
  • text
  • asciidoc
  • rtf
Rows per page
  • 5
  • 10
  • 20
  • 50
  • 100
  • 250
Sort
  • Standard (Relevance)
  • Author A-Ö
  • Author Ö-A
  • Title A-Ö
  • Title Ö-A
  • Publication type A-Ö
  • Publication type Ö-A
  • Issued (Oldest first)
  • Issued (Newest first)
  • Created (Oldest first)
  • Created (Newest first)
  • Last updated (Oldest first)
  • Last updated (Newest first)
  • Disputation date (earliest first)
  • Disputation date (latest first)
  • Standard (Relevance)
  • Author A-Ö
  • Author Ö-A
  • Title A-Ö
  • Title Ö-A
  • Publication type A-Ö
  • Publication type Ö-A
  • Issued (Oldest first)
  • Issued (Newest first)
  • Created (Oldest first)
  • Created (Newest first)
  • Last updated (Oldest first)
  • Last updated (Newest first)
  • Disputation date (earliest first)
  • Disputation date (latest first)
Select
The maximal number of hits you can export is 250. When you want to export more records please use the Create feeds function.
  • 1.
    Alklind Taylor, Anna-Sofia
    et al.
    University of Skövde, School of Humanities and Informatics. University of Skövde, The Informatics Research Centre.
    Backlund, Per
    University of Skövde, School of Humanities and Informatics. University of Skövde, The Informatics Research Centre.
    Niklasson, Lars
    University of Skövde, School of Humanities and Informatics. University of Skövde, The Informatics Research Centre.
    The Coaching Cycle: A Coaching-by-Gaming Approach in Serious Games2012In: Journal Simulation & Gaming, ISSN 1046-8781, E-ISSN 1552-826X, Vol. 43, no 5, p. 648-672Article in journal (Refereed)
    Abstract [en]

    Military organizations have a long history of using simulations, role-play, and games for training. This also encompasses good practices concerning how instructors utilize games and gaming behavior. Unfortunately, the work of instructors is rarely described explicitly in research relating to serious gaming. Decision makers also tend to have overconfidence in the pedagogical power of games and simulations, particularly where the instructor is taken out of the gaming loop. The authors propose a framework, the coaching cycle, that focuses on the roles of instructors. The roles include instructors acting as game players. The fact that the instructors take a more active part in all training activities will further improve learning. The coaching cycle integrates theories of experiential learning (where action precedes theory) and deliberate practice (where the trainee's skill is constantly challenged by a coach). Incorporating a coaching-by-gaming perspective complicates, but also strengthens, the player-centered design approach to game development in that we need to take into account two different types of players: trainees and instructor. Furthermore, the authors argue that the coaching cycle allows for a shift of focus to a more thorough debriefing, because it implies that learning of theoretical material before simulation/game playing is kept to a minimum. This shift will increase the transfer of knowledge.

  • 2.
    Andler, Sten F.
    et al.
    University of Skövde, School of Humanities and Informatics.
    Niklasson, Lars
    University of Skövde, School of Humanities and Informatics.
    Olsson, Björn
    University of Skövde, School of Humanities and Informatics.
    Persson, Anne
    University of Skövde, School of Humanities and Informatics.
    Planstedt, Tomas
    Ericsson Microwave Systems AB, Skövde, Sweden.
    De Vin, Leo J.
    University of Skövde, School of Technology and Society.
    Wangler, Benkt
    University of Skövde, School of Humanities and Informatics.
    Ziemke, Tom
    University of Skövde, School of Humanities and Informatics.
    Information Fusion from Databases, Sensors and Simulations: A Collaborative Research Program2005In: Proceedings: 29th Annual IEEE/NASA Software Engineering Workshop, IEEE Computer Society, 2005, p. 234-241Conference paper (Refereed)
    Abstract [en]

    This paper provides an overview of a collaborative research program in information fusion from databases, sensors and simulations. Information fusion entails the combination of data from multiple sources, to generate information that cannot be derived from the individual sources. This area is of strategic importance for industry and defense, as well as public administration areas such as health care, and needs to be pursued as an academic subject. A large number of industrial partners are supporting and participating in the development of the area. The paper describes the program’s general approach and main research areas, with a particular focus on the role of information fusion in systems development

  • 3.
    Bodén, Mikael
    et al.
    University of Skövde, School of Humanities and Informatics.
    Niklasson, Lars
    University of Skövde, School of Humanities and Informatics.
    Features of distributed representations for tree-structures: A study of RAAM1995Report (Other academic)
    Abstract [en]

    This paper presents an in-depth analysis of properties of patterns, generated by the Recursive Auto-Associative Memory, based on the idea that representational features can be detected by a classification network. The intension of this analysis is to examine the actual reasons for the success of connectionist processes acting on super-positional activity vectors generated in the fashion described. We show that the structure supplied during training is maintained and is extractable from the generated pattern. Further, we show that the influence of the actual constituents in the structures supplied during training is not necessarily available, in the generated patterns, for holistic processing. The outlook for holistic processing is therefore limited unless new forms can be found which take into account, what Sharkey and Jackson call, `whole net' representations.

  • 4.
    Boström, Henrik
    et al.
    University of Skövde, School of Humanities and Informatics.
    Andler, Sten F.
    University of Skövde, School of Humanities and Informatics.
    Brohede, Marcus
    University of Skövde, School of Humanities and Informatics.
    Johansson, Ronnie
    University of Skövde, School of Humanities and Informatics.
    Karlsson, Alexander
    University of Skövde, School of Humanities and Informatics.
    van Laere, Joeri
    University of Skövde, School of Humanities and Informatics.
    Niklasson, Lars
    University of Skövde, School of Humanities and Informatics.
    Nilsson, Marie
    University of Skövde, School of Humanities and Informatics.
    Persson, Anne
    University of Skövde, School of Humanities and Informatics.
    Ziemke, Tom
    University of Skövde, School of Humanities and Informatics.
    On the Definition of Information Fusion as a Field of Research2007Report (Other academic)
    Abstract [en]

    A more precise definition of the field of information fusion can be of benefit to researchers within the field, who may use uch a definition when motivating their own work and evaluating the contribution of others. Moreover, it can enable researchers and practitioners outside the field to more easily relate their own work to the field and more easily understand the scope of the techniques and methods developed in the field. Previous definitions of information fusion are reviewed from that perspective, including definitions of data and sensor fusion, and their appropriateness as definitions for the entire research field are discussed. Based on strengths and weaknesses of existing definitions, a novel definition is proposed, which is argued to effectively fulfill the requirements that can be put on a definition of information fusion as a field of research.

  • 5.
    Brax, Christoffer
    et al.
    University of Skövde, School of Humanities and Informatics. University of Skövde, The Informatics Research Centre.
    Karlsson, Alexander
    University of Skövde, School of Humanities and Informatics. University of Skövde, The Informatics Research Centre.
    Andler, Sten F.
    University of Skövde, School of Humanities and Informatics. University of Skövde, The Informatics Research Centre.
    Johansson, Ronnie
    University of Skövde, School of Humanities and Informatics. University of Skövde, The Informatics Research Centre.
    Niklasson, Lars
    University of Skövde, School of Humanities and Informatics. University of Skövde, The Informatics Research Centre.
    Evaluating Precise and Imprecise State-Based Anomaly Detectors for Maritime Surveillance2010In: Proceedings of the 13th International Conference on Information Fusion, IEEE conference proceedings, 2010, p. Article number 5711997-Conference paper (Refereed)
    Abstract [en]

    We extend the State-Based Anomaly Detection approach by introducing precise and imprecise anomaly detectors using the Bayesian and credal combination operators, where evidences over time are combined into a joint evidence. We use imprecision in order to represent the sensitivity of the classification regarding an object being  normal or anomalous. We evaluate the detectors on a real-world maritime dataset containing recorded AIS data and show that the anomaly detectors outperform   previously proposed detectors based on Gaussian mixture models and kernel density estimators. We also show that our introduced anomaly detectors perform slightly better than the State-Based Anomaly Detection approach with a sliding window.

  • 6.
    Brax, Christoffer
    et al.
    University of Skövde, School of Humanities and Informatics. University of Skövde, The Informatics Research Centre.
    Laxhammar, Rikard
    University of Skövde, School of Humanities and Informatics.
    Niklasson, Lars
    University of Skövde, School of Humanities and Informatics. University of Skövde, The Informatics Research Centre.
    Approaches for detecting behavioural anomalies in public areas using video surveillance data2008In: Proceedings of SPIE Europe 2008, 16–18 September 2008, Cardiff, Wales, United Kingdom / [ed] David A. Huckridge; Reinhard R. Ebert, SPIE - International Society for Optical Engineering, 2008, p. Article number 711318-Conference paper (Refereed)
  • 7.
    Brax, Christoffer
    et al.
    University of Skövde, School of Humanities and Informatics. University of Skövde, The Informatics Research Centre.
    Niklasson, Lars
    University of Skövde, School of Humanities and Informatics. University of Skövde, The Informatics Research Centre.
    An approach for increased supply chain security by using automatic detection of anomalous vehicle behavior2009In: CD-ROM Proceedings of the 6th International Conference on Modeling Decisions for Artificial Intelligence (MDAI 2009), 2009, p. 165-176Conference paper (Refereed)
    Abstract [en]

    In recent years, the development of low-cost GPS transceivers has made it possible to equip all trucks in a fleet with equipment for automatically reporting the status of the trucks to a fleet management system. The downside is that the huge amount of information that is gathered must be evaluated in real-time by an operator. We propose the use of a data-driven anomaly detection algorithm that learns "normal" vehicle behaviour and detects anomalous behaviour such as smuggling, accidents and hijacking, The algorithm is evaluated on real-world data from trucks and commuters equipped with GPS transceivers. The results give initial support to the claim that anomaly detection based on statistical learning can be used to support human descision making. This ability can increase supply chain security by alerting an operator on anomalous vehicle behaviour.

  • 8.
    Brax, Christoffer
    et al.
    University of Skövde, School of Humanities and Informatics. University of Skövde, The Informatics Research Centre.
    Niklasson, Lars
    University of Skövde, School of Humanities and Informatics. University of Skövde, The Informatics Research Centre.
    Enhanced situational Awareness in the Maritime Domain: An Agent-based Approach for Situation Management2009In: Intelligent Sensing, Situation Management, Impact Assessment, and Cyber-Sensing: Proceedings of SPIE Defense, Security, and Sensing 2009 / [ed] Stephen Mott, John F Buford, Gabriel Jakobson, SPIE Press , 2009, p. Aticle ID 735203-Conference paper (Refereed)
    Abstract [en]

    Maritime Domain Awareness is important for both civilian and military applications. An important part of MDA is detection of unusual vessel activities such as piracy, smuggling, poaching, collisions, etc. Today's interconnected sensorsystems provide us with huge amounts of information over large geographical areas which can make the operators reach their cognitive capacity and start to miss important events. We propose and agent-based situation management system that automatically analyse sensor information to detect unusual activity and anomalies. The system combines knowledge-based detection with data-driven anomaly detection. The system is evaluated using information from both radar and AIS sensors.

  • 9.
    Brax, Christoffer
    et al.
    University of Skövde, School of Humanities and Informatics. University of Skövde, The Informatics Research Centre.
    Niklasson, Lars
    University of Skövde, School of Humanities and Informatics. University of Skövde, The Informatics Research Centre.
    Laxhammar, Rikard
    University of Skövde, School of Humanities and Informatics. University of Skövde, The Informatics Research Centre.
    An ensemble approach for increased anomaly detection performance in video surveillance data2009In: Proceedings of the 12th International Conference on Information Fusion (FUSION 2009), Seattle, Washington, USA, 6–9 July 2009, IEEE conference proceedings, 2009, p. 694-701Conference paper (Refereed)
    Abstract [en]

    The increased societal need for surveillance and the decrease in cost of sensors have led to a number of new challenges. The problem is not to collect data but to use it effectively for decision support. Manual interpretation of huge amounts of data in real-time is not feasible; the operator of a surveillance system needs support to analyze and understand all incoming data. In this paper an approach to intelligent video surveillance is presented, with emphasis on finding behavioural anomalies. Two different anomaly detection methods are compared and combined. The results show that it is possible to best increase the total detection performance by combining two different anomaly detectors rather than employing them independently.

     

  • 10.
    Brax, Christoffer
    et al.
    University of Skövde, School of Humanities and Informatics. University of Skövde, The Informatics Research Centre.
    Niklasson, Lars
    University of Skövde, School of Humanities and Informatics. University of Skövde, The Informatics Research Centre.
    Smedberg, Martin
    Saab Microwave Systems, Saab AB, Gothenburg, Sweden.
    Finding behavioural anomalies in public areas using video surveillance data2008In: Proceedings of the 11th International Conference on Information Fusion, IEEE conference proceedings, 2008, p. 1655-1662Conference paper (Refereed)
    Abstract [en]

     In this paper we propose an approach forvdetecting anomalies in data from visual surveillancevsensors. The approach includes creating a structure for representing data, building “normal models” by filling the structure with data for the situation at hand, and finally detecting deviations in the data. The approach allows detections based on the incorporation of a priori knowledge about the situation and on data-driven analysis. The main advantages with the approach compared to earlier work is the low computational requirements, iterative update of normal models and a high explainability of found anomalies. The proposed approach is evaluated off-line using real-world data and the results support that the approach could be used to detect anomalies in real-time applications.

     

  • 11.
    Buason, Gunnar
    et al.
    University of Skövde, School of Humanities and Informatics.
    Niklasson, Lars
    University of Skövde, School of Humanities and Informatics.
    NFFP3+ Concepts and Methods2004Report (Other academic)
  • 12.
    Dahlbom, Anders
    et al.
    University of Skövde, School of Humanities and Informatics. University of Skövde, The Informatics Research Centre.
    Niklasson, Lars
    University of Skövde, School of Humanities and Informatics. University of Skövde, The Informatics Research Centre.
    Evolving Petri Net Situation Templates for Situation Recognition2009In: Proceedings of the 3rd Skövde Workshop on Information Fusion Topics (SWIFT 2009) / [ed] Ronnie Johansson, Joeri van Laere and Jonas Mellin, University of Skövde , 2009, p. 11-16Conference paper (Refereed)
    Abstract [en]

     

    Situation recognition is an important problem to address in order to enhance the capabilities of modern surveillance systems. Situation recognition is concerned with finding a priori defined situations that possibly are instantiated in the present flow of information. It can be a rather tricky task to manually define templates for situations that evolve over time, and to at the same time achieve good results with respect to recall and precision on a situation recognition task. In this paper we present some initial results concerning the task of applying genetic algorithms to evolve Petri net based situation templates of interesting situations. Our results show that it is possible to evolve Petri nets that are on par with manually defined templates. However, more research is needed in order to establish the actual effects it has on recall and precision.

     

  • 13.
    Dahlbom, Anders
    et al.
    University of Skövde, School of Humanities and Informatics. University of Skövde, The Informatics Research Centre.
    Niklasson, Lars
    University of Skövde, School of Humanities and Informatics. University of Skövde, The Informatics Research Centre.
    Goal-Directed Hierarchical Dynamic Scripting for RTS Games2006In: Proceedings of the Second Artificial Intelligence and Interactive Digital Entertainment Conference (AIIDE -06) / [ed] John Laird & Jonathan Schaeffer, 2006, p. 21-28Conference paper (Refereed)
    Abstract [en]

    Learning how to defeat human players is a challenging task in today’s commercial computer games. This paper suggests a goal-directed hierarchical dynamic scripting approach for incorporating learning into real-time strategy games. Two alternatives for shortening the re-adaptation time when using dynamic scripting are also presented. Finally, this paper presents an effective way of throttling the performance of the adaptive artificial intelligence system. Put together, the approach entails the possibility of an artificial intelligence opponent to be challenging for a human player, but not too challenging.

  • 14.
    Dahlbom, Anders
    et al.
    University of Skövde, School of Humanities and Informatics. University of Skövde, The Informatics Research Centre.
    Niklasson, Lars
    University of Skövde, School of Humanities and Informatics. University of Skövde, The Informatics Research Centre.
    Trajectory Clustering for Coastal Surveillance2007In: 10th International Conference on Information Fusion (FUSION 2007),, IEEE Press, 2007, p. 1-8Conference paper (Refereed)
    Abstract [en]

    Achieving superior situation awareness is a key task for military, as well as civilian, decision makers. Today, automatic systems provide us with an excellent opportunity for assisting the human decision maker in achieving this awareness. Due to the potential of information overload one important aspect is to understand where to focus attention. Anomaly detection is concerned with finding deviations from normalcy and it is an increasingly important topic when providing decision support, since it can give hints towards where more analysis is needed. In this paper we explore trajectory clustering as a means for representing normal behavior in a coastal surveillance scenario. Trajectory clustering however suffers from some drawbacks in this type of setting and we therefore propose a new approach, spline-based clustering, with a potential for solving the task of representing the normal course of events.

  • 15.
    Dahlbom, Anders
    et al.
    University of Skövde, School of Humanities and Informatics. University of Skövde, The Informatics Research Centre.
    Niklasson, Lars
    University of Skövde, School of Humanities and Informatics. University of Skövde, The Informatics Research Centre.
    Falkman, Göran
    University of Skövde, School of Humanities and Informatics. University of Skövde, The Informatics Research Centre.
    A Component-based Simulator for Supporting Research on Situation Recognition2009In: Intelligent Sensing, Situation Management, Impact Assessment, and Cyber-Sensing: Proceedings of SPIE Defense, Security, and Sensing 2009 / [ed] John F Koegel Buford, SPIE - International Society for Optical Engineering, 2009, p. Article ID 735206-Conference paper (Refereed)
    Abstract [en]

    Research on information fusion and situation management within the military domain, is often focused on data-driven approaches for aiding decision makers in achieving situation awareness. We have in a companion paper identified situation recognition as an important topic for further studies on knowledge-driven approaches. When developing new algorithms it is of utmost importance to have data for studying the problem at hand (as well as for evaluation purposes). This often become a problem within the military domain as there is a high level of secrecy, resulting in a lack of data, and instead one often needs to resort to artificial data. Many tools and simulation environments can be used for constructing scenarios in virtual worlds. Most of these are however data-centered, that is, their purpose is to simulate the real-world as accurately as possible, in contrast to simulating complex scenarios. In high-level information fusion we can however often assume that lower-level problems have already been solved - thus the separation of abstraction - and we should instead focus on solving problems concerning complex relationships, i.e. situations and threats. In this paper we discuss requirements that research on situation recognition puts on simulation tools. Based on these requirements we present a component-based simulator for quickly adapting the simulation environment to the needs of the research problem at hand. This is achieved by defining new components that define behaviors of entities in the simulated world.

  • 16.
    Dahlbom, Anders
    et al.
    University of Skövde, School of Humanities and Informatics. University of Skövde, The Informatics Research Centre.
    Niklasson, Lars
    University of Skövde, School of Humanities and Informatics. University of Skövde, The Informatics Research Centre.
    Falkman, Göran
    University of Skövde, School of Humanities and Informatics. University of Skövde, The Informatics Research Centre.
    Attempting to increase the Performance of Petri net based Situation Recognition2010In: Proceedings of the 22nd Benelux Conference on Artificial Intelligence, Benelux Association for Artificial Intelligence , 2010Conference paper (Refereed)
    Abstract [en]

    Situation recognition is an important problem to solve for introducing new capabilities in surveillance applications. It is concerned with recognizing a priori defined situations of interest, which are characterized as being of temporal and concurrent nature. The purpose is to aid decision makers with focusing on information that is known to likely be important for them, given their goals. Besides the two most important problems: knowing what to recognize and being able to recognize it, there are three main problems coupled to real time recognition of situations. Computational complexity — we need to process data and information within bounded time. Tractability — human operators must be able to easily understand what is being modelled. Expressability — we must be able to express situations at suitable levels of abstraction. In this paper we attempt to lower the computational complexity of a Petri net based approach for situation.

  • 17.
    Dahlbom, Anders
    et al.
    University of Skövde, School of Humanities and Informatics. University of Skövde, The Informatics Research Centre.
    Niklasson, Lars
    University of Skövde, School of Humanities and Informatics. University of Skövde, The Informatics Research Centre.
    Falkman, Göran
    University of Skövde, School of Humanities and Informatics. University of Skövde, The Informatics Research Centre.
    DESIRER: a Development Environment for Situation Recognition Research2010In: Proceedings 2010 Second WRI Global Congress on Intelligent Systems: GCIS 2010: Volume 1 / [ed] Xinhan Huang, Li Da Xu, Zu De Zhou, Zhun Fan, M. M. Gupta, & Pan Wang, Los Alamitos, CA: IEEE Computer Society, 2010, p. 143-148Conference paper (Refereed)
    Abstract [en]

    Situation recognition is an important problem within the surveillance domain, which addresses the problem of recognizing a priori defined patterns of interesting situations that may be of concurrent and temporal nature, and which possibly are occurring in the present flow of data and information. There may be many viable approaches, with different properties, for addressing this problem however, something they must have in common is good efficiency and high performance. In order to determine if a potential solution has these properties, it is a necessity to have access to test and development environments. In this paper we present DESIRER, a development environment for working with situation recognition, and for evaluating and comparing different approaches.

  • 18.
    Dahlbom, Anders
    et al.
    University of Skövde, School of Humanities and Informatics. University of Skövde, The Informatics Research Centre.
    Niklasson, Lars
    University of Skövde, School of Humanities and Informatics. University of Skövde, The Informatics Research Centre.
    Falkman, Göran
    University of Skövde, School of Humanities and Informatics. University of Skövde, The Informatics Research Centre.
    Evolving Petri Nets for Situation Recognition2010In: GEM 2010: Proceedings of the 2010 International Conference on Genetic and Evolutionary Methods / [ed] Hamid R. Arabnia, Ray R. Hashemi, Ashu M. G. Solo, CSREA Press, 2010, p. 29-35Conference paper (Refereed)
    Abstract [en]

    Situation recognition is an important problem to address for developing newcapabilities in the surveillance domain.  It is concerned with recognizing a priori defined situations of interest, which can be of concurrent and temporal nature, possibly occurring in a continuous flow of data and information. It is however a complex task to manually define what constitutes an interesting situation, and we therefore investigate the possibility of using genetic algorithms for evolving Petri nets for situation recognition. Our results show that: (1) it is possible to evolve complex Petri nets, (2) it is possible to increase the performance of manually  designed Petri nets, and (3) a dynamic genome representation consisting of  complex genes is beneficial compared to a representation consisting of bit strings.

  • 19.
    Dahlbom, Anders
    et al.
    University of Skövde, School of Humanities and Informatics. University of Skövde, The Informatics Research Centre.
    Niklasson, Lars
    University of Skövde, School of Humanities and Informatics. University of Skövde, The Informatics Research Centre.
    Falkman, Göran
    University of Skövde, School of Informatics. University of Skövde, The Informatics Research Centre.
    Situation Recognition and Hypothesis Management Using Petri Nets2009In: Modeling Decisions for Artificial Intelligence: Proceedings of the 6th International Conference (MDAI 2009) / [ed] Vicenç Torra, Yasuo Narukawa, Masahiro Inuiguchi, Springer Berlin/Heidelberg, 2009, p. 303-314Conference paper (Refereed)
    Abstract [en]

    Situation recognition – the task of tracking states and identifying situations - is a problem that is important to look into for aiding decision makers in achieving enhanced situation awareness. The purpose of situation recognition is, in contrast to producing more data and information, to aid decision makers in focusing on information that is important for them, i.e. to detect potentially interesting situations. In this paper we explore the applicability of a Petri net based approach for modeling and recognizing situations, as well as for managing the hypothesis space coupled to matching situation templates with the present stream of data.

     

  • 20.
    Dahlbom, Anders
    et al.
    University of Skövde, School of Humanities and Informatics. University of Skövde, The Informatics Research Centre.
    Niklasson, Lars
    University of Skövde, School of Humanities and Informatics. University of Skövde, The Informatics Research Centre.
    Falkman, Göran
    University of Skövde, School of Humanities and Informatics. University of Skövde, The Informatics Research Centre.
    Loutfi, Amy
    Örebro University.
    Towards Template-based Situation Recognition2009In: Intelligent Sensing, Situation Management, Impact Assessment, and Cyber-Sensing: Proceedings of SPIE Defense, Security, and Sensing 2009 / [ed] Stephen Mott, John F. Buford, Gabe Jakobson, Michael J. Mendenhall, SPIE - International Society for Optical Engineering, 2009, p. Article ID 735205-Conference paper (Refereed)
    Abstract [en]

    The process of tracking and identifying developing situations is an ability of importance within the surveillance domain. We refer to this as situation recognition and believe that it can enhance situation awareness for decision makers. Situation recognition requires that many subproblems are solved. For instance, we need to establish which situations are interesting, how to represent these situations, and which inferable events and states that can be used for representing them. We also need to know how to track and identify situations and how to determine the correlation between present information about situations with knowledge. For some of these subproblems, data-driven approaches are suitable, whilst knowledge-driven approaches are more suitable for others. In this paper we discuss our current research efforts and goals concerning template-based situation recognition. We provide a categorization of approaches for situation recognition together with a formalization of the template-based situation recognition problem. We also discuss this formalization in the light of a pick-pocket scenario. Finally, we discuss future directions for our research on situation recognition. We conclude that situation recognition is an important problem to look into for enhancing the overall situation awareness of decision makers.

  • 21.
    Erlandsson, Tina
    et al.
    Department of Decision Support and Autonomy, Saab AB, Sweden.
    Helldin, Tove
    University of Skövde, School of Humanities and Informatics. University of Skövde, The Informatics Research Centre.
    Falkman, Göran
    University of Skövde, School of Humanities and Informatics. University of Skövde, The Informatics Research Centre.
    Niklasson, Lars
    University of Skövde, School of Humanities and Informatics. University of Skövde, The Informatics Research Centre.
    Information Fusion supporting Team Situation Awareness for Future Fighting Aircraft2010In: FUSION 2010: 13th international Conference on Information Fusion, 26-29 July 2010, EICC, Edinburgh, UK, IEEE conference proceedings, 2010, p. Article number 5712064-Conference paper (Refereed)
    Abstract [en]

    In the military aviation domain, the decisionmaker, i.e. the pilot, often has to process huge amounts of information in order to make correct decisions. This is further aggravated by factors such as time-pressure, high workload and the presence of uncertain information. A support system that aids the pilot to achieve his/her goals has long been considered vital for performance progress in military aviation. Research programs within the domain have studied such support systems, though focus has not been on team collaboration. Based on identified challenges of assessing team situation awareness we suggest an approach to future military aviation support systems based on information fusion. In contrast to most previous work in this area, focus is on supporting team situation awareness, including team threat evaluation. To deal with these challenges, we propose the development of a situational adapting system, which presents information and recommendations based on the current situation.

  • 22.
    Erlandsson, Tina
    et al.
    Aeronautics, Saab AB, Sweden.
    Niklasson, Lars
    University of Skövde, School of Informatics. University of Skövde, The Informatics Research Centre.
    An Air-to-Ground Combat Survivability Model2015In: Journal of Defense Modeling and Simulation: Applications, Methodology, Technology, ISSN 1557-380X, Vol. 12, no 3, p. 273-287Article in journal (Refereed)
    Abstract [en]

    A survivability model can be a useful component of a tactical support system able to aid fighter pilots to assess the risk of getting hit by enemy fire from ground-based threats. This work identifies three desirable properties of such a model: it should allow for evaluating actions; it should enable domain experts to incorporate their knowledge; and it should represent uncertainties both regarding the locations of the threats as well as their future actions. A survivability model issuggested, which calculates the probability that the aircraft can fly a route unharmed and allows for routes of different lengths to be compared. A domain expert can describe the threats by specifying the risk of getting hit at a position of the route without having to consider the earlier actions of the aircraft and the threats. Three different threat models are suggested and compared. The influence of uncertainties regarding the positions of the threats is studied by calculating the probability density function for the survivability. Different representations that take into account both the uncertainty regarding the present and future situation are discussed. The results indicate that the suggested survivability model could be a useful component of a future tactical support system, even though some further development is needed.

  • 23.
    Erlandsson, Tina
    et al.
    Sensor Fusion and Tactical Control, Aeronautics, Saab AB, Sweden.
    Niklasson, Lars
    University of Skövde, School of Informatics. University of Skövde, The Informatics Research Centre.
    Automatic evaluation of air mission routes with respect to combat survival2014In: Information Fusion, ISSN 1566-2535, E-ISSN 1872-6305, Vol. 20, p. 88-98Article in journal (Refereed)
    Abstract [en]

    Aircraft flying in hostile environments are exposed to ground-based air defense systems. It is not always possible to both accomplish the mission and fly outside the range of the enemy's weapon systems, especially if the positions of the enemy's systems are not perfectly known. Automatic evaluation of mission routes from a combat survival perspective could therefore aid the pilots to plan their missions. When updated information regarding the positions and capabilities of the enemy's systems is received during flight, the route could be re-evaluated and the mission could be re-planed or aborted if it is assessed to be too dangerous. The survivability model presented here describes the relation between the aircraft and the enemy's defense systems. It calculates the probabilities that the aircraft is in certain modes along the route, e.g., undetected, tracked or hit. Contrary to previous work, the model is able to capture that the enemy's systems can communicate and that the enemy must track the aircraft before firing a weapon. The survivability model is used to calculate an expected cost for the mission route. The expected cost has the attractive properties of summarizing the route into a single value and is able to take the pilot's risk attitude for the mission into account. The evaluation of the route is influenced by uncertainty regarding the locations of the enemy's sensors and weapons. Monte Carlo simulations are used to capture this uncertainty by calculating the mean and standard deviation for the expected cost. These two parameters give the pilots an assessment of the danger associated with the route as well as the reliability of this assessment. The paper concludes that evaluating routes with the survivability model and the expected cost could aid the pilots to plan and execute their missions. (C) 2014 Elsevier B.V. All rights reserved.

  • 24.
    Erlandsson, Tina
    et al.
    Aeronautics, SAAB AB, Linköping, Sweden.
    Niklasson, Lars
    University of Skövde, School of Humanities and Informatics. University of Skövde, The Informatics Research Centre.
    Calculating Uncertainties in Situation Analysis for Fighter Aircraft Combat Survivability2012In: Proceedings of the 15th International Conference on Information Fusion (FUSION 2012), IEEE Computer Society, 2012, p. 196-203Conference paper (Refereed)
    Abstract [en]

    The aim of situation analysis is to assess the relevant objects in the surroundings and interpret their relations and their impact in order for a decision maker to achieve situation awareness and be able to make suitable decisions. However, the information regarding the relevant objects is typically uncertain, which will induce uncertainty in the result from the situation analysis. If the kinematic states of the objects are estimated with a tracking filter, the estimates can be considered as random variables. Furthermore, the situation analysis algorithm is a function of these estimates entailing that the result from the situation analysis is random variable. This paper studies the fighter aircraft domain and a situation analysis algorithm that calculates the combat survivability, i.e., the probability that the aircraft can a fly a route inside hostile territory without getting hit by enemy fire. The survivability of different routes can be compared in order to decide where to fly. However, the uncertainties regarding the threats' positions imply that the survivability is uncertain and can be desribed as a random variable with a distribution. The unscented transform (UT) is here used for calculating the mean and standard deviation (std) of the survivability in a few scenarios with threats located on the ground. Simulations show that the position uncertainties affect both the mean and std of the survivability and that UT gives similar estimates as a Monte Carlo (MC) approach. UT therefore seems to be a promising approach for calculating the uncertainty in the survivability, which is more computational efficient than MC.

  • 25.
    Erlandsson, Tina
    et al.
    Saab Aeronautics, Linköping, Sweden.
    Niklasson, Lars
    University of Skövde, School of Humanities and Informatics. University of Skövde, The Informatics Research Centre.
    Comparing Air Mission Routes from a Combat Survival Perspective2013In: Proceedings of the Twenty-Sixth International Florida Artificial Intelligence Research Society Conference / [ed] Chutima Boonthum-Denecke and G. Michael Youngblood, AAAI Press, 2013, p. 58-63Conference paper (Refereed)
    Abstract [en]

    An aircraft flying inside hostile territory is exposed to the risk of getting detected and tracked by the enemy’s sensors,  and  subsequently  hit  by  its  weapons.  This paper  describes  a  combat  survivability  model  that can be used for assessing the risks associated with a mission route. In contrast to previous work, the model describes both the risk of getting tracked and the risk of getting hit, as well as the dependency between these risks.  Three  different  ways  of  using  the  model  for comparing routes from a combat survival perspective are  suggested.  The  survivability  for  the  end  point, i.e., the probability of flying the entire route without getting hit, is a compact way of summarizing the risks. Visualizing  how  the  risks  vary  along  the  route  can be  used  for  identifying  critical  parts  of  the  mission. Finally, assigning weights to different risks allow the opportunity to take preferences regarding risk exposure into account.

  • 26.
    Erlandsson, Tina
    et al.
    Aeronautics, SAAB AB, Linköping, Sweden.
    Niklasson, Lars
    University of Skövde, School of Humanities and Informatics. University of Skövde, The Informatics Research Centre.
    Threat Assessment for Missions in Hostile Territory - From the Aircraft Perspective2013In: Proceedings of the 16th international conference on information fusion (FUSION 2013), IEEE Press, 2013, p. 1856-1862Conference paper (Refereed)
  • 27. Erlandsson, Tina
    et al.
    Niklasson, Lars
    University of Skövde, School of Humanities and Informatics. University of Skövde, The Informatics Research Centre.
    Uncertainty Measures for Sensor Management in a Survivability Application2011In: Informatik 2011 / [ed] Heiss, H-U., Pepper, P., Schlingloff, H. and Schneider, J., Bonner Köller Verlag , 2011Conference paper (Refereed)
    Abstract [en]

    When flying a mission, a fighter pilot is exposed to the risk of being hit by enemy fire. A tactical support system can aid the pilot by calculating the survivability of a given route, which is the probability that the fighter pilot can fly the route with-out being hit. The survivability estimate will be uncertain due to uncertainty in the information about threats in the area. In this paper, we investigate the uncertainty in the estimate of the survivability and compare two different measures of uncertainty; standard deviation and entropy. Furthermore, we discuss how these measures can be used for sensor management and discuss a few issues that need to be addressed in the design of a sensor management system in a fighter aircraft.

  • 28.
    Erlandsson, Tina
    et al.
    Aeronautics/Electronic Defence Systems, SAAB AB, Linköping/Göteborg, Sweden.
    Niklasson, Lars
    University of Skövde, School of Humanities and Informatics. University of Skövde, The Informatics Research Centre.
    Nordlund, Per-Johan
    Aeronautics/Electronic Defence Systems, SAAB AB, Linköping/Göteborg, Sweden.
    Warston, Håkan
    Aeronautics/Electronic Defence Systems, SAAB AB, Linköping/Göteborg, Sweden.
    Modeling Fighter Aircraft Mission Survivability2011In: Proceedings of the 14th International Conference on Information Fusion (FUSION 2011), IEEE conference proceedings, 2011, p. 999-1006Conference paper (Refereed)
    Abstract [en]

    A fighter aircraft flying a mission is often exposed to ground-based threats such as surface-to-air missile (SAM) sites. The fighter pilot needs to take actions to minimize the risk of being shot down, but at the same time be able to accomplish the mission. In this paper we propose a survivability model, which describes the probability that the aircraft will be able to fly a given route without being hit by incoming missiles. Input to this model can consist of sensor measurements collected during flight as well as intelligence data gathered before the mission. This input is by nature uncertain and we therefore investigate the influence of uncertainty in the input to the model. Finally we propose a number of decision support functions that can be developed based on the suggested model such as countermeasure management, mission planning and sensor management.

  • 29.
    Gustavsson, Per M.
    et al.
    Security and Defense Solutions/Training and Simulation, Saab, Sweden.
    Hieb, Michael R.
    Center of Excellence for C4I, George Mason University, United States.
    Moore, Philip
    De Montfort University, United Kingdom.
    Eriksson, Patric
    Gothia Science Park, Skövde, Sweden / De Montfort University, United Kingdom.
    Niklasson, Lars
    University of Skövde, School of Humanities and Informatics. University of Skövde, The Informatics Research Centre.
    Operations Intent and Effects Model2011In: The Journal of Defence Modeling and Simulation: Applications, Methodology, Technology, ISSN 1548-5129, E-ISSN 1557-380X, Vol. 8, no 1, p. 37-59Article in journal (Refereed)
    Abstract [en]

    Military missions in the 21st century are characterized by combinations of traditional symmetric conventional warfare, irregular warfare, and operations other than war. The inherent uncertainty in an actual mission and the variety of potential organizations (e.g. multi-agency, non-governmental, private volunteer, international, international corporations) from several countries that support the mission makes collaboration and co-ordination a key capability for command and control. The ability to communicate and automatically process intent and effects is vital in order for a commander to cooperate with other organizations and agencies and lead subordinates in such a way that the overall mission is completed in the best possible way, including exploitation of fleeting opportunities, i.e. enable for self-synchronization amongst teams and allow for subordinate initiatives. However, intent and effects are often absent in the current and forthcoming digitalized information models, and if intent and effects are present it is likely to be found that the representations are made as free-text fields based on natural language. However, such messages are very difficult to disambiguate, particularly for automated machine systems. The overall objective for the Operations Intent and Effects Model is to support operational and simulated systems by a conceptual intent and effects model and a formalism that is human and machine interpretable.

  • 30.
    Hansson, Andreas
    et al.
    University of Skövde, School of Humanities and Informatics.
    Niklasson, Lars
    University of Skövde, School of Humanities and Informatics.
    Using RAAM to model human sequence representation and processing2001Report (Other academic)
    Abstract [en]

    The human mind has been studied from various perspectives. Some adopt an analytical approach by analyzing humans, yet others attempt to construct automated systems exhibiting certain aspects of mind. Here we argue that connectionist architectures generally fail to exhibit important aspects of mind. We present a number of aspects, relating to human short-term and long-term memory during sequence representation and processing. These aspects are then used as a means to measure the explanatory power of connectionist architectures. We find that connectionist architectures (specifically Recursive Auto-Associative Memories, RAAM) generally fail to model important aspects of the short-term and long-term memory, when representing and processing sequences. Some aspects are correctly modeled, whereas others are modeled incorrectly or it is an open question whether or not they can be modeled at all. From this we go on to present the areas in which more research is needed, before connectionist RAAM-like architectures can be finally claimed to model important aspects of short-term and long-term memory.

  • 31.
    Hansson, Andreas
    et al.
    University of Skövde, School of Humanities and Informatics.
    Niklasson, Lars
    University of Skövde, School of Humanities and Informatics.
    Using Segmentation to Control the Retrieval of Data2006In: The 2006 IEEE International Joint Conference on Neural Network Proceedings, IEEE, 2006, p. 1764-1769Conference paper (Other academic)
    Abstract [en]

    One problem when storing sequential data using recurrent neural networks is that it is hard to preserve long term dependencies. Only the most recently stored data tend to be accurately recalled. One approach for reducing this recency effect has been to divide the data into segments and store the segments separately. This approach has provided promising results in prediction and classification domains. This paper analyzes in what way recall of the stored data is affected by segmentation. It is concluded that segmentation enables the control of which data that can be recalled. The problem of preserving long term dependencies in recurrent neural networks can therefore be reduced.

  • 32.
    Helldin, Tove
    et al.
    University of Skövde, School of Humanities and Informatics. University of Skövde, The Informatics Research Centre.
    Erlandsson, Tina
    Department of Data Fusion and Tactical Control, Saab AB, Linköping, Sweden.
    Niklasson, Lars
    University of Skövde, School of Humanities and Informatics. University of Skövde, The Informatics Research Centre.
    Falkman, Göran
    University of Skövde, School of Humanities and Informatics. University of Skövde, The Informatics Research Centre.
    Situational Adapting System supporting Team Situation Awareness2010In: Unmanned/Unattended Sensors and Sensor Networks VII: Proceedings of SPIE Security & Defence 2010 / [ed] Edward M. Carapezza, SPIE - International Society for Optical Engineering, 2010, p. Article No. 78330S-Conference paper (Refereed)
    Abstract [en]

    Military fighter pilots have to make suitable decisions fast in an environment where continuously increasing flows of information from sensors, team members and databases are provided. Not only do the huge amounts of data aggravate the pilots’ decision making process: time-pressure, presence of uncertain data and high workload are factors that can worsen the performance of pilot decision making. In this paper, initial ideas of how to support the pilots accomplishing their tasks are presented. Results from interviews with two fighter pilots are described as well as a discussion about how these results can guide the design of a military fighter pilot decision support system, with focus on team cooperation.

  • 33.
    Johansson, Ulf
    et al.
    University of Borås.
    König, Ralph
    University of Borås.
    Niklasson, Lars
    University of Skövde, School of Humanities and Informatics. University of Skövde, The Informatics Research Centre.
    Inconsistency: Friend or Foe2007In: Proceedings of International Joint Conference on Neural Networks, Orlando, Florida, USA, August 12-17, 2007, IEEE Press, 2007, p. 1383-1388Conference paper (Refereed)
    Abstract [en]

    One way of obtaining accurate yet comprehensible models is to extract rules from opaque predictive models. When evaluating rule extraction algorithms, one frequently used criterion is consistency; i.e. the algorithm must produce similar rules every time it is applied to the same problem. Rule extraction algorithms based on evolutionary algorithms are, however, inherently inconsistent, something that is regarded as their main drawback. In this paper, we argue that consistency is an overvalued criterion, and that inconsistency can even be beneficial in some situations. The study contains two experiments, both using publicly available data sets, where rules are extracted from neural network ensembles. In the first experiment, it is shown that it is normally possible to extract several different rule sets from an opaque model, all having high and similar accuracy. The implication is that consistency in that perspective is useless; why should one specific rule set be considered superior? Clearly, it should instead be regarded as an advantage to obtain several accurate and comprehensible descriptions of the relationship. In the second experiment, rule extraction is used for probability estimation. More specifically, an ensemble of extracted trees is used in order to obtain probability estimates. Here, it is exactly the inconsistency of the rule extraction algorithm that makes the suggested approach possible.

  • 34.
    Johansson, Ulf
    et al.
    University of Borås, Sch Business & Informat, Sweden.
    König, Rikard
    University of Borås, Sch Business & Informat, Sweden.
    Löfström, Tuve
    Univerity off Borås, Sch Business & Informat, Sweden.
    Niklasson, Lars
    University of Skövde, School of Humanities and Informatics. University of Skövde, The Informatics Research Centre.
    Increasing Rule Extraction Accuracy by Post-processing GP Trees2008In: IEEE Congress on Evolutionary Computation, IEEE Press, 2008, p. 3010-3015Conference paper (Refereed)
    Abstract [en]

    Genetic programming (GP), is a very general and efficient technique, often capable of outperforming more specialized techniques on a variety of tasks. In this paper, we suggest a straightforward novel algorithm for post-processing of GP classification trees. The algorithm iteratively, one node at a time, searches for possible modifications that would result in higher accuracy. More specifically, the algorithm for each split evaluates every possible constant value and chooses the best. With this design, the post-processing algorithm can only increase training accuracy, never decrease it. In this study, we apply the suggested algorithm to GP trees, extracted from neural network ensembles. Experimentation, using 22 UCI datasets, shows that the post-processing results in higher test set accuracies on a large majority of datasets. As a matter of fact, for two setups of three evaluated, the increase in accuracy is statistically significant.

  • 35.
    Johansson, Ulf
    et al.
    School of Business and Informatics, University of Borås, Sweden.
    König, Rikard
    School of Business and Informatics, University of Borås, Sweden.
    Löfström, Tuve
    School of Business and Informatics, University of Borås, Sweden.
    Niklasson, Lars
    University of Skövde, School of Humanities and Informatics. University of Skövde, The Informatics Research Centre.
    Using Imaginary Ensembles to Select GP Classifiers2010In: Genetic Programming: 13th European Conference, EuroGP 2010, Istanbul, Turkey, April 7-9, 2010. Proceedings / [ed] Anna Isabel Esparcia-Alcázar, Anikó Ekárt, Sara Silva, Stephen Dignum, A. Şima Uyar, Springer Berlin/Heidelberg, 2010, p. 278-288Conference paper (Refereed)
    Abstract [en]

    When predictive modeling requires comprehensible models, most data miners will use specialized techniques producing rule sets or decision trees. This study, however, shows that genetically evolved decision trees may very well outperform the more specialized techniques. The proposed approach evolves a number of decision trees and then uses one of several suggested selection strategies to pick one specific tree from that pool. The inherent inconsistency of evolution makes it possible to evolve each tree using all data, and still obtain somewhat different models. The main idea is to use these quite accurate and slightly diverse trees to form an imaginary ensemble, which is then used as a guide when selecting one specific tree. Simply put, the tree classifying the largest number of instances identically to the ensemble is chosen. In the experimentation, using 25 UCI data sets, two selection strategies obtained significantly higher accuracy than the standard rule inducer J48.

  • 36.
    Johansson, Ulf
    et al.
    School of Business and Informatics, University of Borås, Borås, Sweden.
    König, Rikard
    School of Business and Informatics, University of Borås, Borås, Sweden.
    Löfström, Tuve
    School of Business and Informatics, University of Borås, Borås, Sweden.
    Sönströd, Cecilia
    School of Business and Informatics, University of Borås, Borås, Sweden.
    Niklasson, Lars
    University of Skövde, School of Humanities and Informatics. University of Skövde, The Informatics Research Centre.
    Post-processing Evolved Decision Trees2009In: Studies in Computational Intelligence, ISSN 1860-949X, E-ISSN 1860-9503, Vol. 204, p. 149-164Article in journal (Other academic)
    Abstract [en]

    Although Genetic Programming (GP) is a very general technique, it is also quite powerful. As a matter of fact, GP has often been shown to outperform more specialized techniques on a variety of tasks. In data mining, GP has successfully been applied to most major tasks; e.g. classification, regression and clustering. In this chapter, we introduce, describe and evaluate a straightforward novel algorithm for post-processing genetically evolved decision trees. The algorithm works by iteratively, one node at a time, search for possible modifications that will result in higher accuracy. More specifically, the algorithm, for each interior test, evaluates every possible split for the current attribute and chooses the best. With this design, the post-processing algorithm can only increase training accuracy, never decrease it. In the experiments, the suggested algorithm is applied to GP decision trees, either induced directly from datasets, or extracted from neural network ensembles. The experimentation, using 22 UCI datasets, shows that the suggested post-processing technique results in higher test set accuracies on a large majority of the datasets. As a matter of fact, the increase in test accuracy is statistically significant for one of the four evaluated setups, and substantial on two out of the other three.

  • 37.
    Johansson, Ulf
    et al.
    School of Business and Informatics, University of Borås, Sweden.
    König, Rikard
    School of Business and Informatics, University of Borås, Sweden.
    Niklasson, Lars
    University of Skövde, School of Humanities and Informatics. University of Skövde, The Informatics Research Centre.
    Automatically Balancing Accuracy and Comprehensibility in Predictive Modeling2005In: 2005 8th International Conference on Information Fusion (FUSION) Philadelphia, PA 25-28 July 2005: Volume 2 of 2, IEEE conference proceedings, 2005, p. 1554-1560Conference paper (Refereed)
    Abstract [en]

    One specific problem, when performing predictive modeling, is the tradeoff between accuracy and comprehensibility. When comprehensible models are required this normally rules out high-accuracy techniques like neural networks and committee machines. Therefore, an automated choice of a standard technique, known to generally produce sufficiently accurate and comprehensible models, would be of great value. In this paper it is argued that this requirement is met by an ensemble of classifiers, followed by rule extraction. The proposed technique is demonstrated, using an ensemble of common classifiers and our rule extraction algorithm G-REX, on 17 publicly available data sets. The results presented demonstrate that the suggested technique performs very well. More specifically, the ensemble clearly outperforms the individual classifiers regarding accuracy, while the extracted models have accuracy similar to the individual classifiers. The extracted models are, however, significantly more compact than corresponding models created directly from the data set using he standard tool CART; thus providing higher comprehensibility.

  • 38.
    Johansson, Ulf
    et al.
    School of Business and Informatics, University of Borås, Borås, Sweden.
    König, Rikard
    School of Business and Informatics, University of Borås, Borås, Sweden.
    Niklasson, Lars
    University of Skövde, School of Humanities and Informatics. University of Skövde, The Informatics Research Centre.
    Genetic rule extraction optimizing brier score2010In: Proceedings of the 12th Annual Genetic and Evolutionary Computation Conference, GECCO '10 / [ed] Pelikan, Martin & Branke, Jürgen, New York: Association for Computing Machinery (ACM), 2010, p. 1007-1014Conference paper (Refereed)
    Abstract [en]

    Most highly accurate predictive modeling techniques produce opaque models. When comprehensible models are required, rule extraction is sometimes used to generate a transparent model, based on the opaque. Naturally, the extracted model should be as similar as possible to the opaque. This criterion, called fidelity, is therefore a key part of the optimization function in most rule extracting algorithms. To the best of our knowledge, all existing rule extraction algorithms targeting fidelity use 0/1 fidelity, i.e., maximize the number of identical classifications. In this paper, we suggest and evaluate a rule extraction algorithm utilizing a more informed fidelity criterion. More specifically, the novel algorithm, which is based on genetic programming, minimizes the difference in probability estimates between the extracted and the opaque models, by using the generalized Brier score as fitness function. Experimental results from 26 UCI data sets show that the suggested algorithm obtained considerably higher accuracy and significantly better AUC than both the exact same rule extraction algorithm maximizing 0/1 fidelity, and the standard tree inducer J48. Somewhat surprisingly, rule extraction using the more informed fidelity metric normally resulted in less complex models, making sure that the improved predictive performance was not achieved on the expense of comprehensibility. Copyright 2010 ACM.

  • 39.
    Johansson, Ulf
    et al.
    School of Business and Informatics, University of Borås, Borås, Sweden.
    König, Rikard
    School of Business and Informatics, University of Borås, Borås, Sweden.
    Niklasson, Lars
    University of Skövde, School of Humanities and Informatics. University of Skövde, The Informatics Research Centre.
    Genetically Evolved kNN Ensembles2009In: Data Mining: Special Issue in Annals of Information Systems / [ed] Robert Stahlbock, Sven F. Crone, Stefan Lessmann, Springer Science+Business Media B.V., 2009, 1, p. 299-313Chapter in book (Other academic)
    Abstract [en]

    Both theory and a wealth of empirical studies have established that ensembles are more accurate than single predictive models. For the ensemble approach to work, base classifiers must not only be accurate but also diverse, i.e., they should commit their errors on different instances. Instance-based learners are, however, very robust with respect to variations of a data set, so standard resampling methods will normally produce only limited diversity. Because of this, instance-based learners are rarely used as base classifiers in ensembles. In this chapter, we introduce a method where genetic programming is used to generate kNN base classifiers with optimized k-values and feature weights. Due to the inherent inconsistency in genetic programming (i.e., different runs using identical data and parameters will still produce different solutions) a group of independently evolved base classifiers tend to be not only accurate but also diverse. In the experimentation, using 30 data sets from the UCI repository, two slightly different versions of kNN ensembles are shown to significantly outperform both the corresponding base classifiers and standard kNN with optimized k-values, with respect to accuracy and AUC.

  • 40.
    Johansson, Ulf
    et al.
    University of Skövde, School of Humanities and Informatics. Department of Business and Informatics, University of Borås, Sweden.
    König, Rikard
    University of Skövde, School of Humanities and Informatics. Department of Business and Informatics, University of Borås, Sweden.
    Niklasson, Lars
    University of Skövde, School of Humanities and Informatics.
    The Truth is in There: Rule Extraction from Opaque Models Using Genetic Programming2004In: Proceedings of the Seventeenth International Florida Artificial Intelligence Research Society Conference, FLAIRS 2004 / [ed] Valerie Barr, Zdravko Markov, AAAI Press, 2004, p. 658-663Conference paper (Other academic)
  • 41. Johansson, Ulf
    et al.
    Löfström, Tove
    University of Skövde, School of Humanities and Informatics.
    König, Richard
    University of Skövde, School of Humanities and Informatics.
    Sönströd, Cecilia
    Niklasson, Lars
    University of Skövde, School of Humanities and Informatics.
    Rule Extraction from Opaque Models: A Slightly Different Perspective2006In: 6th International Conference on Machine Learning and Applications, IEEE Computer Society, 2006, p. 22-27Conference paper (Refereed)
    Abstract [en]

    When performing predictive modeling, the key criterion is always accuracy. With this in mind, complex techniques like neural networks or ensembles are normally used, resulting in opaque models impossible to interpret. When models need to be comprehensible, accuracy is often sacrificed by using simpler techniques directly producing transparent models; a tradeoff termed the accuracy vs. comprehensibility tradeoff. In order to reduce this tradeoff, the opaque model can be transformed into another, interpretable, model; an activity termed rule extraction. In this paper, it is argued that rule extraction algorithms should gain from using oracle data; i.e. test set instances, together with corresponding predictions from the opaque model. The experiments, using 17 publicly available data sets, clearly show that rules extracted using only oracle data were significantly more accurate than both rules extracted by the same algorithm, using training data, and standard decision tree algorithms. In addition, the same rules were also significantly more compact; thus providing better comprehensibility. The overall implication is that rules extracted in this fashion will explain the predictions made on novel data better than rules extracted in the standard way; i.e. using training data only.

  • 42. Johansson, Ulf
    et al.
    Löfström, Tuve
    König, Rikard
    Niklasson, Lars
    University of Skövde, School of Humanities and Informatics. University of Skövde, The Informatics Research Centre.
    Accurate Neural Network Ensembles Using Genetic Programming2006In: Proceedings of SAIS: The 23rd Annual Workshop of the Swedish Artificial Intelligence Society, Swedish Artificial Intelligence Society - SAIS, Umeå universitet , 2006Conference paper (Other academic)
    Abstract [en]

    Abstract: In this paper we present and evaluate a novel algorithm for ensemble creation. The main idea of the algorithm is to first independently train a fixed number of neural networks (here ten) and then use genetic programming to combine these networks into an ensemble. The use of genetic programming makes it possible to not only consider ensembles of different sizes, but also to use ensembles as intermediate building blocks. The final result is therefore more correctly described as an ensemble of neural network ensembles. The experiments show that the proposed method, when evaluated on 22 publicly available data sets, obtains very high accuracy, clearly outperforming the other methods evaluated. In this study several micro techniques are used, and we believe that they all contribute to the increased performance.

    One such micro technique, aimed at reducing overtraining, is the training method, called tombola training, used during genetic evolution. When using tombola training, training data is regularly resampled into new parts, called training groups. Each ensemble is then evaluated on every training group and the actual fitness is determined solely from the result on the hardest part.

  • 43. Johansson, Ulf
    et al.
    Löfström, Tuve
    König, Rikard
    Niklasson, Lars
    University of Skövde, School of Humanities and Informatics. University of Skövde, The Informatics Research Centre.
    Building Neural Network Ensembles using Genetic Programming2006In: The International Joint Conference on Neural Networks 2006, IEEE Press, 2006, p. 2239-2244Conference paper (Refereed)
    Abstract [en]

    algorithm for ensemble creation. The main idea of the algorithm is to first independently train a fixed number of neural networks (here ten) and then use genetic programming to combine these networks into an ensemble. The use of genetic programming makes it possible to not only consider ensembles of different sizes, but also to use ensembles as intermediate building blocks. The final result is therefore more correctly described as an ensemble of neural network ensembles. The experiments show that the proposed method, when evaluated on 22 publicly available data sets, obtains very high accuracy, clearly outperforming the other methods evaluated. In this study several micro techniques are used, and we believe that they all contribute to the increased performance. One such micro technique, aimed at reducing overtraining, is the training method, called tombola training, used during genetic evolution. When using tombola training, training data is regularly resampled into new parts, called training groups. Each ensemble is then evaluated on every training group and the actual fitness is determined solely from the result on the hardest part.

  • 44.
    Johansson, Ulf
    et al.
    School of Business and Informatics, University of Borås, Sweden.
    Löfström, Tuve
    School of Business and Informatics, University of Borås, Sweden.
    König, Rikard
    School of Business and Informatics, University of Borås, Sweden.
    Niklasson, Lars
    University of Skövde, School of Humanities and Informatics. University of Skövde, The Informatics Research Centre.
    Genetically Evolved Trees Representing Ensembles2006In: Artificial intelligence and soft computing - ICAISC 2006: 8th international conference, Zakopane, Poland, June 25 - 29, 2006 ; proceedings, 2006, p. 613-622Conference paper (Refereed)
    Abstract [en]

    We have recently proposed a novel algorithm for ensemble creation called GEMS (Genetic Ensemble Member Selection). GEMS first trains a fixed number of neural networks (here twenty) and then uses genetic programming to combine these networks into an ensemble. The use of genetic programming makes it possible for GEMS to not only consider ensembles of different sizes, but also to use ensembles as intermediate building blocks. In this paper, which is the first extensive study of GEMS, the representation language is extended to include tests partitioning the data, further increasing flexibility. In addition, several micro techniques are applied to reduce overfitting, which appears to be the main problem for this powerful algorithm. The experiments show that GEMS, when evaluated on 15 publicly available data sets, obtains very high accuracy, clearly outperforming both straightforward ensemble designs and standard decision tree algorithms.

  • 45. Johansson, Ulf
    et al.
    Löfström, Tuve
    König, Rikard
    Niklasson, Lars
    University of Skövde, School of Humanities and Informatics. University of Skövde, The Informatics Research Centre.
    Introducing GEMS * a Novel Technique for Ensemble Creation2006In: 19th Florida Artificial Intelligence Research Society Conference (FLAIRS) 06, AAAI Press, 2006, p. 700-705Conference paper (Refereed)
    Abstract [en]

    The main contribution of this paper is to suggest a novel technique for automatic creation of accurate ensembles. The technique proposed, named GEMS, first trains a large number of neural networks (here either 20 or 50) and then uses genetic programming to build the ensemble by combining available networks. The use of genetic programming makes it possible for GEMS to not only consider ensembles of very different sizes, but also to use ensembles as intermediate building blocks which could be further combined into larger ensembles. To evaluate the performance, GEMS is compared to different ensembles where networks are selected based on individual test set accuracy. The experiments use four publicly available data sets and the results are very promising for GEMS. On two data sets, GEMS has significantly higher accuracy than the competing ensembles, while there is no significant difference on the other two.

  • 46.
    Johansson, Ulf
    et al.
    University of Skövde, School of Humanities and Informatics. School of Business and Informatics, University of Borås, Borås, Sweden.
    Löfström, Tuve
    School of Business and Informatics, University of Borås.
    König, Rikard
    School of Business and Informatics, University of Borås.
    Niklasson, Lars
    University of Skövde, School of Humanities and Informatics.
    Why Not Use an Oracle When You Got One?2006In: Neural Information Processing: Letters and Reviews, ISSN 1738-2572, Vol. 10, no 8-9, p. 227-236Article in journal (Refereed)
    Abstract [en]

    The primary goal of predictive modeling is to achieve high accuracy when the model is applied to novel data. For certain problems this requires the use of complex techniques like neural networks or ensembles, resulting in opaque models that are hard or impossible to interpret. For some domains this is unacceptable, since models need to be comprehensible. To achieve comprehensibility, accuracy is often sacrificed by using simpler techniques; a tradeoff termed the accuracy vs. comprehensibility tradeoff. Another, frequently studied, alternative is rule extraction; i.e. the activity where another, transparent, model is generated from the opaque model. In this paper it is argued that existing rule extraction algorithms do not use all information available, and typically should benefit from also using oracle data; i.e. test set instances, together with corresponding predictions from the opaque model. The experiments, using fifteen publicly available data sets, clearly show that rules extracted using either just oracle data or training data augmented with oracle data, will explain the predictions significantly better than rules extracted in the standard way; i.e. using training data only.

  • 47. Johansson, Ulf
    et al.
    Löfström, Tuve
    Niklasson, Lars
    University of Skövde, School of Humanities and Informatics. University of Skövde, The Informatics Research Centre.
    Accuracy on a Hold-out Set: The Red Herring of Data Mining2006In: 23rd Annual Workshop of the Swedish Artificial Intelligence Society, Swedish Artificial Intelligence Society - SAIS, Umeå universitet , 2006Conference paper (Other academic)
    Abstract [en]

    Abstract: When performing predictive modeling, the overall goal is to generate models likely to have high accuracy when applied to novel data. A technique commonly used to maximize generalization accuracy is to create ensembles of models, e.g., averaging the output from a number of individual models. Several, more or less sophisticated techniques, aimed at either directly creating ensembles or selecting ensemble members from a pool of available models, have been suggested. Many techniques utilize a part of the available data not used for the training of the models (a hold-out set) to rank and select either ensembles or ensemble members based on accuracy on that set. The obvious underlying assumption is that increased accuracy on the hold-out set is a good indicator of increased generalization capability on novel data. Or, put in another way, that there is high correlation between accuracy on the hold-out set and accuracy on yet novel data. The experiments in this study, however, show that this is generally not the case; i.e. there is little to gain from selecting ensembles using hold-out set accuracy. The experiments also show that this low correlation holds for individual neural networks as well; making the entire use of hold-out sets to compare predictive models questionable

  • 48.
    Johansson, Ulf
    et al.
    Department of Business and Informatics, University of Borås, Borås, Sweden .
    Löfström, Tuve
    University of Skövde, School of Humanities and Informatics. University of Skövde, The Informatics Research Centre.
    Niklasson, Lars
    University of Skövde, School of Humanities and Informatics. University of Skövde, The Informatics Research Centre.
    Evaluating Standard Techniques for Implicit Diversity2008In: Advances in Knowledge Discovery and Data Mining: 12th Pacific-Asia Conference PAKDD 2008 / [ed] Washio, T; Suzuki, E; Ting, KM; Inokuchi, A, Springer Berlin/Heidelberg, 2008, p. 592-599Conference paper (Refereed)
    Abstract [en]

    When performing predictive modeling, ensembles are often utilized in order to boost accuracy. The problem of how to maximize ensemble accuracy is, however, far from solved. In particular, the relationship between ensemble diversity and accuracy is, especially for classification, not completely understood. More specifically, the fact that ensemble diversity and base classifier accuracy are highly correlated, makes it necessary to balance these properties instead of just maximizing diversity. In this study, three standard techniques to obtain implicit diversity in neural network ensembles are evaluated using 14 UCI data sets. The experiments show that standard resampling; i.e. dividing the training data by instances, produces more diverse models, but at the expense of base classifier accuracy, thus resulting in less accurate ensembles. Building ensembles using neural networks with heterogeneous architectures improves test set accuracies, but without actually increasing the diversity. The results regarding resampling using features are inconclusive, the ensembles become more diverse, but the level of test set accuracies is unchanged. For the setups evaluated, ensemble training accuracy and base classifier training accuracy are positively correlated with ensemble test accuracy, but the opposite holds for diversity; i.e. ensembles with low diversity are generally more accurate.

  • 49.
    Johansson, Ulf
    et al.
    School of Business and Informatics, University of Borås, SE-501 90 Borås, Sweden.
    Löfström, Tuve
    School of Business and Informatics, University of Borås, SE-501 90 Borås, Sweden.
    Niklasson, Lars
    University of Skövde, School of Humanities and Informatics. University of Skövde, The Informatics Research Centre.
    The Importance of Diversity in Neural Network Ensembles: An Empirical Investigation2007In: The International Joint Conference on Neural Networks, IEEE Press, 2007, p. 661-666Conference paper (Refereed)
    Abstract [en]

    When designing ensembles, it is almost an axiom that the base classifiers must be diverse in order for the ensemble to generalize well. Unfortunately, there is no clear definition of the key term diversity, leading to several diversity measures and many, more or less ad hoc, methods for diversity creation in ensembles. In addition, no specific diversity measure has shown to have a high correlation with test set accuracy. The purpose of this paper is to empirically evaluate ten different diversity measures, using neural network ensembles and 11 publicly available data sets. The main result is that all diversity measures evaluated, in this study too, show low or very low correlation with test set accuracy. Having said that, two measures; double fault and difficulty show slightly higher correlations compared to the other measures. The study furthermore shows that the correlation between accuracy measured on training or validation data and test set accuracy also is rather low. These results challenge ensemble design techniques where diversity is explicitly maximized or where ensemble accuracy on a hold-out set is used for optimization.

  • 50.
    Johansson, Ulf
    et al.
    School of Business and Informatics, University of Borås, Sweden.
    Niklasson, Lars
    University of Skövde, School of Humanities and Informatics. University of Skövde, The Informatics Research Centre.
    Evolving Decision Trees Using Oracle Guides2009In: 2009 IEEE Symposium on Computational Intelligence and Data Mining (CIDM 2009) Proceedings, IEEE, 2009, p. 238-244Conference paper (Refereed)
    Abstract [en]

    Some data mining problems require predictive models to be not only accurate but also comprehensible. Comprehensibility enables human inspection and understanding of the model, making it possible to trace why individual predictions are made. Since most high-accuracy techniques produce opaque models, accuracy is, in practice, regularly sacrificed for comprehensibility. One frequently studied technique, often able to reduce this accuracy vs. comprehensibility tradeoff, is rule extraction, i.e., the activity where another, transparent, model is generated from the opaque. In this paper, it is argued that techniques producing transparent models, either directly from the dataset, or from an opaque model, could benefit from using an oracle guide. In the experiments, genetic programming is used to evolve decision trees, and a neural network ensemble is used as the oracle guide. More specifically, the datasets used by the genetic programming when evolving the decision trees, consist of several different combinations of the original training data and "oracle data", i.e., training or test data instances, together with corresponding predictions from the oracle. In total, seven different ways of combining regular training data with oracle data were evaluated, and the results, obtained on 26 UCI datasets, clearly show that the use of an oracle guide improved the performance. As a matter of fact, trees evolved using training data only had the worst test set accuracy of all setups evaluated. Furthermore, statistical tests show that two setups, both using the oracle guide, produced significantly more accurate trees, compared to the setup using training data only.

12 1 - 50 of 74
CiteExportLink to result list
Permanent link
Cite
Citation style
  • apa
  • harvard1
  • ieee
  • modern-language-association-8th-edition
  • vancouver
  • Other style
More styles
Language
  • de-DE
  • en-GB
  • en-US
  • fi-FI
  • nn-NO
  • nn-NB
  • sv-SE
  • Other locale
More languages
Output format
  • html
  • text
  • asciidoc
  • rtf