We propose a new area to focus upon as a research challenge: How to use personalized technology t... more We propose a new area to focus upon as a research challenge: How to use personalized technology to connect our cultural heritage experiences to our "daily" lives. In particular we want to connect the visitor's museum experiences with outside relevant mobile cultural heritage experiences. We begin to examine what processes and technologies will be necessary to accomplish the when, what and how of this challenge.
ContextOver the years, several modeling languages for requirements have been proposed. These lang... more ContextOver the years, several modeling languages for requirements have been proposed. These languages employ different conceptual approaches, including scenario-based and goal-oriented ones. Empirical studies providing evidence about requirements model comprehensibility are rare, especially when addressing languages that belong to different modeling approaches.Objective This work aims to compare the comprehensibility of requirements models expressed in different but comparable modeling approaches from a requirement analysts’ perspective. In particular, in this paper we compare the comprehensibility of requirements models expressed in two visual modeling languages: Use Case, which is scenario-based, and Tropos, which exploits goal-oriented modeling. We further compare the effort required for comprehending the different models, and the derived productivity in each case.Method Requirements model comprehensibility is measured here in the context of three types of tasks that analysts us...
Filtering rules may refer to various attributes of the data items subject to the filtering proces... more Filtering rules may refer to various attributes of the data items subject to the filtering process. In personal rulebased filtering systems, each user has his/her own personal filtering rules. In stereotype rule-based filtering systems, a user is assigned to a group of similar users (his/her stereotype) from which he/she inherits the stereotype's filtering profile. This study compares the effectiveness of the two alternative rule-based filtering methods: stereotype-based rules versus personal rules. We conducted a comparison between filtering effectiveness when using the personal rules or when using the stereotype-based rules. Although, intuitively, personal filtering rules seem to be more effective because each user has his own tailored rules, our comparative study reveals that stereotype filtering rules yield more effective results. We believe that this is because users find it difficult to evaluate their filtering preferences accurately. The results imply that by using a stereotype it is possible not only to overcome the problem of user effort required to generate a manual rule-based profile, but at the same time even provide a better initial user profile.
Information filtering (IF) systems usually filter data items by correlating a vector of terms (ke... more Information filtering (IF) systems usually filter data items by correlating a vector of terms (keywords) that represent the user profile with similar vectors of terms that represent the data items (e.g. documents). The terms that represent the data items can be determined by (human) experts (e.g. authors of documents) or by automatic indexing methods. In this study we employ an artificial neural-network (ANN) as an alternative method for both filtering and term selection, and compare its effectiveness to "traditional" methods. In an earlier study we developed and examined the performance of an IF system that employed content-based and stereotypic rule-based filtering methods, in the domain of e-mail messages. In this study we train a large-scale ANN-based filter which uses meaningful terms in the same database of email messages as input, and use it to predict the relevancy of those messages.
Portzel R., Malaka R. "A task based approach for ontology evaluation", Proceedings of ECAI, 2004.... more Portzel R., Malaka R. "A task based approach for ontology evaluation", Proceedings of ECAI, 2004. Reinberger M.M., Spyns P. "Discovering knowledge in text for the learning of DOGMA inspired ontologies". Proceedings of OLP04 (workshop on ontology learning and population at ECAI), 2004. Rigau G., Rodriguez H. and Agirre E. "Building accurate semantic taxonomies from Monolingual MRDs".
The growth of available entertainment information services, such as movies and CD listings, or tr... more The growth of available entertainment information services, such as movies and CD listings, or travels and recreational activities, raises a need for personalization techniques for filtering and adapting contents to customer's interest and needs. Personalization technologies rely on users data, represented as User Models (UMs). UMs built by specific services are usually not transferable due to commercial competition and models' representation heterogeneity. This paper focuses on the second obstacle and discusses architecture for mediating UMs across different domains of entertainment. The mediation facilitates improving the accuracy of the UMs and upgrading the provided personalization.
One of the important challenges for personalised, context-aware information delivery within build... more One of the important challenges for personalised, context-aware information delivery within buildings is to be able to show the user a map with their own location as well as the locations of points of interest for them. One very desirable property of a personalised, context-aware mobile application is that it can operate at the same time as preserving the user"s privacy. To achieve this, we need to do on-device positioning and personalisation. This paper presents the design of a platform, and its implementation for the retrieval of publicly available building data (symbolic maps and associated radio-frequency infrastructure point locations) for the purpose of coarse-grained indoor positioning on mobile devices. In comparison to other indoor positioning systems, this work focuses on the mechanism through which building data is made available to mobile devices, as too the motivation in providing generic coarse-grained indoor positioning based on the use of existing infrastructure and building data.
Proceedings of the 24th International Conference on World Wide Web - WWW '15 Companion, 2015
This paper presents a new method for extracting unique features of items based on their textual r... more This paper presents a new method for extracting unique features of items based on their textual reviews. The method is built of two similar iterations of applying a weighting scheme and then clustering the resultant set of vectors. In the first iteration, restaurants of similar food genres are grouped together into clusters. The second iteration reduces the importance of common terms in each such cluster, and highlights those that are unique to each specific restaurant. Clustering the restaurants again, now according to their unique features, reveals very interesting connections between the restaurants.
For museum curators it is imperative to learn, analyze, and understand the behavior patterns of t... more For museum curators it is imperative to learn, analyze, and understand the behavior patterns of the visitors in their museum. Recent developments in the field of indoor positioning systems make the acquisition and availability of visitor behavior data more attainable. However, the analysis of such data remains a challenge due to its noisiness, complexity and sheer size. The current paper applies information visualization techniques to analyze this data and make it more accessible to museum curators and personnel. We first provide a detailed description of the application domain including an analysis of the curators' information needs and a description of how a dataset on visitors' spatiotemporal behavior could be acquired. In order to address the curators' needs, we designed a visualization to encode and convey the information based on a newly adjusted visual glyph that we call Tangram Diagrams. We thereby focus on the adaptability of the technique to a particular domain, rather than on the novelty aspects of the technique itself. We have evaluated our design decisions empirically, and conducted an expert study to describe the insights gained and the value of the information obtained from the visualization. The contribution of this work is twofold. First, we apply information visualization to the museum domain and discuss how it extends to general indoor spatiotemporal behavior analysis. Second, we show how a visual glyph metaphor can be applied in different ways and contexts to efficiently encode multi-faceted information.
Personalization in the "museum visit" scenario is extremely challenging, especially since in many... more Personalization in the "museum visit" scenario is extremely challenging, especially since in many cases visitors come to the museum for the first, and may be the last time in their life. There is therefore a need to generate an effective user model quickly without any prior knowledge. Furthermore, the initial definition of a user model is also challenging since it should be built in a non-intrusive manner. Understanding visitors' behavioral patterns may help in initializing their user models and supporting them better. This chapter reports three stages of analysis of behavior patterns of museum visitors. The first step assesses, following past ethnographic research, whether a distinct stereotype of behavior can be identified; the second shows that visitors' behavior is not always consistent; the third shows that, in spite of the inconsistency, prediction of visitor type, is possible.
Document categorization is a daily task in every organization, but it is a very subjective proces... more Document categorization is a daily task in every organization, but it is a very subjective process. While automatic document categorization has been widely studied, much challenging research still remains to support user subjective categorization. This study evaluates and compares the application of Self-Organizing Maps (SOM) and Learning Vector Quantization (LVQ) to automatic document classification according to a subjectively predefined set of clusters in a specific domain, and assesses the effect of user inconsistency on this process. Results show that despite the subjective and inconsistent nature of human categorization, automatic document clustering methods correlate well with subjective, personal clustering. Moreover, adapting a system to its users is limited by users' inconsistency, meaning that a perfect adaptation is an impractical goal.
... June 2009 ... Konstan University of Minnesota, USA Tsvi Kuflik University of Haifa, Israel He... more ... June 2009 ... Konstan University of Minnesota, USA Tsvi Kuflik University of Haifa, Israel Henry Lieberman MIT, USA Frank Linton The MITRE Corporation, USA Chee-Kit Looi Nanyang Technological ... 331 Tobias Ley, Barbara Kump, Antonia Maas, Neil Maiden, and Dietrich Albert ...
Providing accurate personalized information services to the users requires knowing their interest... more Providing accurate personalized information services to the users requires knowing their interests and needs, as defined by their User Models (UMs). Since the quality of the personalization depends on the richness of the UMs, services would benefit from enriching their UMs through importing and aggregating partial UMs built by other services from relatively similar domains. The obvious question is how to determine the similarity of domains? This paper proposes to compute inter-domain similarities by exploiting well-known Information Retrieval techniques for comparing textual contents of the Web-sites, classified under the domain nodes in Web-directories. Initial experiments validate feasibility of the proposed approach and raise open research questions.
We propose a new area to focus upon as a research challenge: How to use personalized technology t... more We propose a new area to focus upon as a research challenge: How to use personalized technology to connect our cultural heritage experiences to our "daily" lives. In particular we want to connect the visitor's museum experiences with outside relevant mobile cultural heritage experiences. We begin to examine what processes and technologies will be necessary to accomplish the when, what and how of this challenge.
ContextOver the years, several modeling languages for requirements have been proposed. These lang... more ContextOver the years, several modeling languages for requirements have been proposed. These languages employ different conceptual approaches, including scenario-based and goal-oriented ones. Empirical studies providing evidence about requirements model comprehensibility are rare, especially when addressing languages that belong to different modeling approaches.Objective This work aims to compare the comprehensibility of requirements models expressed in different but comparable modeling approaches from a requirement analysts’ perspective. In particular, in this paper we compare the comprehensibility of requirements models expressed in two visual modeling languages: Use Case, which is scenario-based, and Tropos, which exploits goal-oriented modeling. We further compare the effort required for comprehending the different models, and the derived productivity in each case.Method Requirements model comprehensibility is measured here in the context of three types of tasks that analysts us...
Filtering rules may refer to various attributes of the data items subject to the filtering proces... more Filtering rules may refer to various attributes of the data items subject to the filtering process. In personal rulebased filtering systems, each user has his/her own personal filtering rules. In stereotype rule-based filtering systems, a user is assigned to a group of similar users (his/her stereotype) from which he/she inherits the stereotype's filtering profile. This study compares the effectiveness of the two alternative rule-based filtering methods: stereotype-based rules versus personal rules. We conducted a comparison between filtering effectiveness when using the personal rules or when using the stereotype-based rules. Although, intuitively, personal filtering rules seem to be more effective because each user has his own tailored rules, our comparative study reveals that stereotype filtering rules yield more effective results. We believe that this is because users find it difficult to evaluate their filtering preferences accurately. The results imply that by using a stereotype it is possible not only to overcome the problem of user effort required to generate a manual rule-based profile, but at the same time even provide a better initial user profile.
Information filtering (IF) systems usually filter data items by correlating a vector of terms (ke... more Information filtering (IF) systems usually filter data items by correlating a vector of terms (keywords) that represent the user profile with similar vectors of terms that represent the data items (e.g. documents). The terms that represent the data items can be determined by (human) experts (e.g. authors of documents) or by automatic indexing methods. In this study we employ an artificial neural-network (ANN) as an alternative method for both filtering and term selection, and compare its effectiveness to "traditional" methods. In an earlier study we developed and examined the performance of an IF system that employed content-based and stereotypic rule-based filtering methods, in the domain of e-mail messages. In this study we train a large-scale ANN-based filter which uses meaningful terms in the same database of email messages as input, and use it to predict the relevancy of those messages.
Portzel R., Malaka R. "A task based approach for ontology evaluation", Proceedings of ECAI, 2004.... more Portzel R., Malaka R. "A task based approach for ontology evaluation", Proceedings of ECAI, 2004. Reinberger M.M., Spyns P. "Discovering knowledge in text for the learning of DOGMA inspired ontologies". Proceedings of OLP04 (workshop on ontology learning and population at ECAI), 2004. Rigau G., Rodriguez H. and Agirre E. "Building accurate semantic taxonomies from Monolingual MRDs".
The growth of available entertainment information services, such as movies and CD listings, or tr... more The growth of available entertainment information services, such as movies and CD listings, or travels and recreational activities, raises a need for personalization techniques for filtering and adapting contents to customer's interest and needs. Personalization technologies rely on users data, represented as User Models (UMs). UMs built by specific services are usually not transferable due to commercial competition and models' representation heterogeneity. This paper focuses on the second obstacle and discusses architecture for mediating UMs across different domains of entertainment. The mediation facilitates improving the accuracy of the UMs and upgrading the provided personalization.
One of the important challenges for personalised, context-aware information delivery within build... more One of the important challenges for personalised, context-aware information delivery within buildings is to be able to show the user a map with their own location as well as the locations of points of interest for them. One very desirable property of a personalised, context-aware mobile application is that it can operate at the same time as preserving the user"s privacy. To achieve this, we need to do on-device positioning and personalisation. This paper presents the design of a platform, and its implementation for the retrieval of publicly available building data (symbolic maps and associated radio-frequency infrastructure point locations) for the purpose of coarse-grained indoor positioning on mobile devices. In comparison to other indoor positioning systems, this work focuses on the mechanism through which building data is made available to mobile devices, as too the motivation in providing generic coarse-grained indoor positioning based on the use of existing infrastructure and building data.
Proceedings of the 24th International Conference on World Wide Web - WWW '15 Companion, 2015
This paper presents a new method for extracting unique features of items based on their textual r... more This paper presents a new method for extracting unique features of items based on their textual reviews. The method is built of two similar iterations of applying a weighting scheme and then clustering the resultant set of vectors. In the first iteration, restaurants of similar food genres are grouped together into clusters. The second iteration reduces the importance of common terms in each such cluster, and highlights those that are unique to each specific restaurant. Clustering the restaurants again, now according to their unique features, reveals very interesting connections between the restaurants.
For museum curators it is imperative to learn, analyze, and understand the behavior patterns of t... more For museum curators it is imperative to learn, analyze, and understand the behavior patterns of the visitors in their museum. Recent developments in the field of indoor positioning systems make the acquisition and availability of visitor behavior data more attainable. However, the analysis of such data remains a challenge due to its noisiness, complexity and sheer size. The current paper applies information visualization techniques to analyze this data and make it more accessible to museum curators and personnel. We first provide a detailed description of the application domain including an analysis of the curators' information needs and a description of how a dataset on visitors' spatiotemporal behavior could be acquired. In order to address the curators' needs, we designed a visualization to encode and convey the information based on a newly adjusted visual glyph that we call Tangram Diagrams. We thereby focus on the adaptability of the technique to a particular domain, rather than on the novelty aspects of the technique itself. We have evaluated our design decisions empirically, and conducted an expert study to describe the insights gained and the value of the information obtained from the visualization. The contribution of this work is twofold. First, we apply information visualization to the museum domain and discuss how it extends to general indoor spatiotemporal behavior analysis. Second, we show how a visual glyph metaphor can be applied in different ways and contexts to efficiently encode multi-faceted information.
Personalization in the "museum visit" scenario is extremely challenging, especially since in many... more Personalization in the "museum visit" scenario is extremely challenging, especially since in many cases visitors come to the museum for the first, and may be the last time in their life. There is therefore a need to generate an effective user model quickly without any prior knowledge. Furthermore, the initial definition of a user model is also challenging since it should be built in a non-intrusive manner. Understanding visitors' behavioral patterns may help in initializing their user models and supporting them better. This chapter reports three stages of analysis of behavior patterns of museum visitors. The first step assesses, following past ethnographic research, whether a distinct stereotype of behavior can be identified; the second shows that visitors' behavior is not always consistent; the third shows that, in spite of the inconsistency, prediction of visitor type, is possible.
Document categorization is a daily task in every organization, but it is a very subjective proces... more Document categorization is a daily task in every organization, but it is a very subjective process. While automatic document categorization has been widely studied, much challenging research still remains to support user subjective categorization. This study evaluates and compares the application of Self-Organizing Maps (SOM) and Learning Vector Quantization (LVQ) to automatic document classification according to a subjectively predefined set of clusters in a specific domain, and assesses the effect of user inconsistency on this process. Results show that despite the subjective and inconsistent nature of human categorization, automatic document clustering methods correlate well with subjective, personal clustering. Moreover, adapting a system to its users is limited by users' inconsistency, meaning that a perfect adaptation is an impractical goal.
... June 2009 ... Konstan University of Minnesota, USA Tsvi Kuflik University of Haifa, Israel He... more ... June 2009 ... Konstan University of Minnesota, USA Tsvi Kuflik University of Haifa, Israel Henry Lieberman MIT, USA Frank Linton The MITRE Corporation, USA Chee-Kit Looi Nanyang Technological ... 331 Tobias Ley, Barbara Kump, Antonia Maas, Neil Maiden, and Dietrich Albert ...
Providing accurate personalized information services to the users requires knowing their interest... more Providing accurate personalized information services to the users requires knowing their interests and needs, as defined by their User Models (UMs). Since the quality of the personalization depends on the richness of the UMs, services would benefit from enriching their UMs through importing and aggregating partial UMs built by other services from relatively similar domains. The obvious question is how to determine the similarity of domains? This paper proposes to compute inter-domain similarities by exploiting well-known Information Retrieval techniques for comparing textual contents of the Web-sites, classified under the domain nodes in Web-directories. Initial experiments validate feasibility of the proposed approach and raise open research questions.
Uploads
Papers by Tsvi Kuflik