This document discusses using cross-domain user preferences and personality traits to improve collaborative filtering recommendations. It proposes using personality-based collaborative filtering heuristics and factor models that incorporate both preferences from other domains and personality data. Experiments on a dataset of Facebook likes and personality tests show these methods can enhance recommendations, especially in cold-start situations, by exploiting relationships between preferences and personality across domains like movies, music and books.
In this presentation we present a novel context-aware mobile recommender system for places of interest (POIs). Unlike existing systems, which learn users' preferences solely from their past ratings, it considers also their personality - using the Five Factor Model. Personality is acquired by asking users to complete a brief and entertaining questionnaire as part of the registration process, and is then exploited in: (1) an active learning module that actively acquires ratings-in-context for POIs that users are likely to have experienced, hence reducing the stress and annoyance to rate (or skip rating) items that the users don’t know; and (2) in the recommendation model that builds up on matrix factorization and therefore can be trained even if the users haven’t rated any items yet.
Benjamin Heitmann, PhD defence talk: An Open Framework for Multi-source, Cro...Benjamin Heitmann
The work in this thesis addresses the new challenges and opportunities for online personalisation posed by the emergence of new infrastructures for sharing user preferences and for access to open repositories of data. As a result of these new infrastructures, user profiles can now include data from multiple sources about preferences in multiple domains. This new kind of user profile data requires a cross-domain personalisation approach. However, current cross-domain personalisation approaches are restricted to proprietary social networking ecosystems.
The main problem that we address in this thesis, is to enable cross-domain recommendations without the use of proprietary and closed infrastructure. Towards this goal, we propose an open framework for cross-domain personalisation. Our framework consists of two parts: a conceptual architecture for recommender systems, and our cross-domain personalisation approach. The main enabling technology for our framework is Linked Open Data, as it provides a common data presentation for user preferences and cross-domain links between concepts from many different domains.
As part of our framework, we first propose a conceptual architecture for Linked Open Data recommender systems that provides guidelines and best practices for the typical high level components required for providing personalisation in open ecosystems using Linked Open Data. The architecture has a strong empirical founding, as it based on an empirical survey of 124 RDF-based applications.
Then we introduce and throughly evaluate SemStim, an unsupervised, graph-based algorithm for cross-domain personalisation. It leverages multi-source, domain-neutral user profiles and the semantic network of DBpedia in order to generate recommendations for different source and target domains. The results of our evaluation show that SemStim is able to provide cross-domain recommendations, without any overlap between target and source domains and without using any ratings in the target domain.
We show how we instantiate our proposed conceptual architecture for a prototype implementation that is the outcome of the ADVANSSE collaboration project with CISCO Galway. The prototype shows how to implement our framework for a real-world use case and data.
Our open framework for cross-domain personalisation provides an alternative to existing proprietary cross-domain personalisation approaches. As such, it opens up the potential for novel and innovative personalised services without the risk of user lock-in and data silos.
This document summarizes Pasquale Lops' presentation on semantics-aware content-based recommender systems. It discusses how content-based recommender systems traditionally rely on keyword profiles but have limitations due to issues like multi-word concepts, synonymy, and polysemy. The presentation proposes leveraging semantic text analytics techniques like word sense disambiguation, explicit semantic analysis using Wikipedia, and linked open data to move beyond keyword profiles and address issues like overspecialization and lack of serendipity in recommendations.
Cold-Start Management with Cross-Domain Collaborative Filtering and TagsMatthias Braunhofer
Recommender systems suffer from the new user problem, i.e., the difficulty to make accurate predictions for users that have rated only few items. Moreover, they usually compute recommendations for items just in one domain, such as movies, music, or books. In this paper we deal with such a cold-start situation exploiting cross-domain recommendation techniques, i.e., we suggest items to a user in one target domain by using ratings of other users in a, completely disjoint, auxiliary domain. We present three rating prediction models that make use of information about how users tag items in an auxiliary domain, and how these tags correlate with the ratings to improve the rating prediction task in a different target domain. We show that the proposed techniques can effectively deal with the considered cold-start situation, given that the tags used in the two domains overlap.
Matrix Factorization In Recommender SystemsYONG ZHENG
The document discusses matrix factorization techniques for recommender systems. It begins with an overview of recommender systems and their use of matrix factorization for dimensionality reduction. Principal component analysis and singular value decomposition are described as early linear algebra techniques used for this purpose. The document then focuses on how these techniques evolved into basic and extended matrix factorization methods in recommender systems, using the Netflix Prize competition as an example.
Peter Brusilovsky presents research on user control in adaptive information access systems. The document discusses three types of adaptive systems - adaptive hypermedia, adaptive search, and recommender systems. It explores how each system currently handles user control and collaboration with AI, and proposes methods to improve user control and transparency. These include allowing users to control personalization parameters, fusion of multiple ranking sources, and visualizing recommendation results to better understand the reasoning process. The goal is to develop systems where AI provides information and users make informed decisions, with the human firmly in control.
The document discusses collaborative information retrieval (CIR). It defines CIR as involving groups searching for information together to address a shared task or information need. The document outlines key aspects of collaboration, including tasks like travel planning and bibliographic search. It also discusses how CIR differs from personalized IR, collaborative filtering, and social IR in aspects like the user, personalization, collaboration nature, and information goals. CIR aims to support groups jointly searching for, discussing, and synthesizing information through both synchronous and asynchronous collaboration and communication.
Tutorial on User Profiling with Graph Neural Networks and Related Beyond-Acc...Erasmo Purificato
Slide of the Tutorial on "User Profiling with Graph Neural Networks and Related Beyond-Accuracy Perspectives" @ UMAP'23: 31st ACM Conference on User Modeling, Adaptation and Personalization (June 26, 2023 | Limassol, Cyprus)
In this presentation we present a novel context-aware mobile recommender system for places of interest (POIs). Unlike existing systems, which learn users' preferences solely from their past ratings, it considers also their personality - using the Five Factor Model. Personality is acquired by asking users to complete a brief and entertaining questionnaire as part of the registration process, and is then exploited in: (1) an active learning module that actively acquires ratings-in-context for POIs that users are likely to have experienced, hence reducing the stress and annoyance to rate (or skip rating) items that the users don’t know; and (2) in the recommendation model that builds up on matrix factorization and therefore can be trained even if the users haven’t rated any items yet.
Benjamin Heitmann, PhD defence talk: An Open Framework for Multi-source, Cro...Benjamin Heitmann
The work in this thesis addresses the new challenges and opportunities for online personalisation posed by the emergence of new infrastructures for sharing user preferences and for access to open repositories of data. As a result of these new infrastructures, user profiles can now include data from multiple sources about preferences in multiple domains. This new kind of user profile data requires a cross-domain personalisation approach. However, current cross-domain personalisation approaches are restricted to proprietary social networking ecosystems.
The main problem that we address in this thesis, is to enable cross-domain recommendations without the use of proprietary and closed infrastructure. Towards this goal, we propose an open framework for cross-domain personalisation. Our framework consists of two parts: a conceptual architecture for recommender systems, and our cross-domain personalisation approach. The main enabling technology for our framework is Linked Open Data, as it provides a common data presentation for user preferences and cross-domain links between concepts from many different domains.
As part of our framework, we first propose a conceptual architecture for Linked Open Data recommender systems that provides guidelines and best practices for the typical high level components required for providing personalisation in open ecosystems using Linked Open Data. The architecture has a strong empirical founding, as it based on an empirical survey of 124 RDF-based applications.
Then we introduce and throughly evaluate SemStim, an unsupervised, graph-based algorithm for cross-domain personalisation. It leverages multi-source, domain-neutral user profiles and the semantic network of DBpedia in order to generate recommendations for different source and target domains. The results of our evaluation show that SemStim is able to provide cross-domain recommendations, without any overlap between target and source domains and without using any ratings in the target domain.
We show how we instantiate our proposed conceptual architecture for a prototype implementation that is the outcome of the ADVANSSE collaboration project with CISCO Galway. The prototype shows how to implement our framework for a real-world use case and data.
Our open framework for cross-domain personalisation provides an alternative to existing proprietary cross-domain personalisation approaches. As such, it opens up the potential for novel and innovative personalised services without the risk of user lock-in and data silos.
This document summarizes Pasquale Lops' presentation on semantics-aware content-based recommender systems. It discusses how content-based recommender systems traditionally rely on keyword profiles but have limitations due to issues like multi-word concepts, synonymy, and polysemy. The presentation proposes leveraging semantic text analytics techniques like word sense disambiguation, explicit semantic analysis using Wikipedia, and linked open data to move beyond keyword profiles and address issues like overspecialization and lack of serendipity in recommendations.
Cold-Start Management with Cross-Domain Collaborative Filtering and TagsMatthias Braunhofer
Recommender systems suffer from the new user problem, i.e., the difficulty to make accurate predictions for users that have rated only few items. Moreover, they usually compute recommendations for items just in one domain, such as movies, music, or books. In this paper we deal with such a cold-start situation exploiting cross-domain recommendation techniques, i.e., we suggest items to a user in one target domain by using ratings of other users in a, completely disjoint, auxiliary domain. We present three rating prediction models that make use of information about how users tag items in an auxiliary domain, and how these tags correlate with the ratings to improve the rating prediction task in a different target domain. We show that the proposed techniques can effectively deal with the considered cold-start situation, given that the tags used in the two domains overlap.
Matrix Factorization In Recommender SystemsYONG ZHENG
The document discusses matrix factorization techniques for recommender systems. It begins with an overview of recommender systems and their use of matrix factorization for dimensionality reduction. Principal component analysis and singular value decomposition are described as early linear algebra techniques used for this purpose. The document then focuses on how these techniques evolved into basic and extended matrix factorization methods in recommender systems, using the Netflix Prize competition as an example.
Peter Brusilovsky presents research on user control in adaptive information access systems. The document discusses three types of adaptive systems - adaptive hypermedia, adaptive search, and recommender systems. It explores how each system currently handles user control and collaboration with AI, and proposes methods to improve user control and transparency. These include allowing users to control personalization parameters, fusion of multiple ranking sources, and visualizing recommendation results to better understand the reasoning process. The goal is to develop systems where AI provides information and users make informed decisions, with the human firmly in control.
The document discusses collaborative information retrieval (CIR). It defines CIR as involving groups searching for information together to address a shared task or information need. The document outlines key aspects of collaboration, including tasks like travel planning and bibliographic search. It also discusses how CIR differs from personalized IR, collaborative filtering, and social IR in aspects like the user, personalization, collaboration nature, and information goals. CIR aims to support groups jointly searching for, discussing, and synthesizing information through both synchronous and asynchronous collaboration and communication.
Tutorial on User Profiling with Graph Neural Networks and Related Beyond-Acc...Erasmo Purificato
Slide of the Tutorial on "User Profiling with Graph Neural Networks and Related Beyond-Accuracy Perspectives" @ UMAP'23: 31st ACM Conference on User Modeling, Adaptation and Personalization (June 26, 2023 | Limassol, Cyprus)
Seffah iess11 keynote the human side of service scienceIESS
This keynote talk discusses the need for a human-centric quality model for services and service systems. It motivates moving beyond traditional software and service quality models to account for new factors like trust, privacy, and universality. The talk proposes a quality model with three components: factors, criteria to measure sub-factors, and qualitative/quantitative measures. It suggests using personas, scenarios, and design patterns as tools to model user experiences and derive service systems. The quality model can then assess design artifacts and final services/systems to predict quality in use. The talk aims to bridge technical and human aspects of service design.
This document discusses personas and how they are used as a design tool in interaction design. It provides information on what personas are, how they are developed based on research, their typical elements, and different types of personas. Examples of personas are also provided at the end to illustrate how personas describe fictional people based on research in order to represent different user groups and their goals. The document aims to explain the purpose and process of developing personas to inform interaction design.
Leveraging Graph Neural Networks for User Profiling: Recent Advances and Open...Erasmo Purificato
Slide of the tutorial entitled "Leveraging Graph Neural Networks for User Profiling: Recent Advances and Open Challenges" held at CIKM'23: 32nd ACM International Conference on Information and Knowledge Management (October 21, 2023 | Birmingham, United Kingdom)
This document summarizes a presentation given by Katrien Verbert on interactive recommender systems. It provides an overview of Verbert's research group at KU Leuven, which focuses on recommender systems, visualization, and intelligent user interfaces. The presentation describes various techniques for building interactive recommender systems, including explaining recommendations to users, enabling user interaction with the recommendation process, and addressing challenges like diversity, cold starts, and context awareness. It also summarizes several studies conducted by Verbert and collaborators on interactive music and research talk recommender systems.
We the humans are surrounded with immense unprecedented wealth of information which are available as documents, database or other resources. The access to this information is difficult as by having the information it is not necessary that it could be searched or extracted by the activity we are using. The search engines available should be also customized to handle such queries, sometime the search engines are also not aware of the information they have within the system. The method known as keyword extraction and clustering is introduced which answers this shortcoming by spontaneously recommending documents that are related to users’ current activities. When the communication takes place the important text can be extracted from the conversation and the words extracted are grouped and then are matched with the parts in the document. This method uses Natural Language Processing for extracting of keywords and making the subgroup that is a meaningful statement from the group, another method used is the Hierarchical Clustering for creating clusters form the keywords, here the similarity of two keywords is measured using the Euclidean distance. This paper reviews the various methods for the system.
Redesigning the Open Access Institutional RepositoryEdward Luca
This lecture presents a redesign project of UTS's institutional repository, OPUS. It explains some of the challenges faced by libraries in ensuring eRepository participation, and investigates three user groups - academics, librarians, and information seekers. User experience principles are used to address issues around navigation, terminology, and visual identity.
Presented as a guest lecture to Designing for the Web (Spring 2016) students.
Interactive recommender systems: opening up the “black box”Katrien Verbert
This document summarizes a presentation given by Katrien Verbert on interactive recommender systems. It discusses how recommender systems are typically "black boxes" that do not explain their recommendations to users. The presentation aims to open up this black box by exploring ways to increase transparency, user control, and interaction with recommender systems. Examples of interactive recommender systems that allow users to explore the recommendation process and provide explanations are described. Research on developing and evaluating such interactive systems through multiple user studies is summarized. The objective is to enhance user trust and engagement with recommender systems.
This document discusses recommender systems and approaches used at Netflix. It covers collaborative filtering using user-user and item-item methods, content-based recommendations using item attributes, and hybrid approaches. It provides examples of how Netflix uses collaborative filtering to generate personalized genre rows and social recommendations. Netflix combines many data sources and machine learning models to power its highly personalized recommendation engine.
This document summarizes Katrien Verbert's presentation on interactive recommender systems. The presentation covered several topics:
1) Different types of recommendation techniques including collaborative filtering, content-based filtering, and knowledge-based filtering.
2) Research on interactive recommender systems that aim to increase transparency, user control, and diversity of recommendations.
3) Several user studies conducted on interactive recommender systems that explored talks and conferences, finding that explanations and various levels of user control can impact user experience.
This document summarizes quality criteria that have been proposed for mixed methods research. It discusses quality standards for quantitative and qualitative research separately, and notes the lack of agreement on criteria for qualitative research. The document then examines several proposed sets of criteria for mixed methods research, focusing on appropriate research design, rigorous data collection and analysis, and thorough integration of quantitative and qualitative components. It concludes by describing two studies that sought views from social scientists on quality evaluation in mixed methods studies.
This thesis proposes designing and developing a personalized country recommender system. It outlines introducing the problem motivation and research questions. The document then reviews the state of the art on recommender systems including definitions, data sources, approaches (collaborative filtering, content-based filtering, hybrid filtering), and evaluation metrics. It describes the methodology which includes collecting a training dataset, implementing recommender algorithms (SVD, KNN, co-clustering), and system design. The results and evaluation of the system are then presented.
This thesis examines the digital competence of pre-service and in-service teachers in Anhui Province, China. It includes 4 chapters that 1) review relevant frameworks and literature, 2) assess teachers' digital competence using a validated questionnaire, 3) analyze the results, and 4) will formulate proposals for improving teachers' digital skills. The assessment found teachers' digital competence was at a medium level, with some differences between pre-service and in-service groups. Factors like age, education level, and experience influenced skills. Overall, the study evaluated teachers' abilities and will make recommendations to enhance digital pedagogy in Anhui.
This document provides an overview of personas in libraries and how they can be used. It discusses creating personas for different user groups like public library patrons, students, faculty, and internal library staff. The document also covers how to develop personas through qualitative research methods like interviews and surveys, and how to validate personas with quantitative data from website analytics and surveys. Personas can then be used to help make strategic decisions, design services, and test usability within libraries.
Contextual model of recommending resources on an academic networking portalcsandit
Artificial Intelligence techniques have been instrumental in helping users to handle the large
amount of information on the Internet. The idea of recommendation systems, custom search
engines, and intelligent software has been widely accepted among users who seek assistance in
searching, sorting, classifying, filtering and sharing this vast quantity of information. In this
paper, we present a contextual model of recommendation engine which keeping in mind the
context and activities of a user, recommends resources in an academic networking portal. The
proposed method uses the implicit method of feedback and the concepts relationship hierarchy
to determine the similarity between a user and the resources in the portal. The proposed
algorithm has been tested on an academic networking portal and the results are convincing.
CONTEXTUAL MODEL OF RECOMMENDING RESOURCES ON AN ACADEMIC NETWORKING PORTALcscpconf
Artificial Intelligence techniques have been instrumental in helping users to handle the large amount of information on the Internet. The idea of recommendation systems, custom search engines, and intelligent software has been widely accepted among users who seek assistance insearching, sorting, classifying, filtering and sharing this vast quantity of information. In thispaper, we present a contextual model of recommendation engine which keeping in mind the context and activities of a user, recommends resources in an academic networking portal. Theproposed method uses the implicit method of feedback and the concepts relationship hierarchy to determine the similarity between a user and the resources in the portal. The proposed algorithm has been tested on an academic networking portal and the results are convincing
A Literature Review of Quantitative Persona CreationMinjoon Kim
This paper provides a literature review of quantitative persona creation (QPC), which uses algorithmic methods to create personas from numerical and textual data. The paper analyzes 47 research papers on QPC. It identifies three stages of evolution for QPC methods, from using basic clustering techniques on survey data to leveraging large datasets from social media and APIs. Key trends include higher automation, interactive persona systems, and combining automatic and manual methods. The paper also discusses research gaps in developing standards, addressing ethics concerns, and retaining benefits of qualitative persona creation. It concludes with recommendations to advance the field of QPC.
Human-centered AI: how can we support lay users to understand AI?Katrien Verbert
The document summarizes research on human-centered AI and how to support lay users in understanding AI. It discusses various research projects that aim to explain model outcomes to increase user trust and acceptance. It explores how personal characteristics like need for cognition can impact the effectiveness of explanations. The research also looks at different application domains for AI like healthcare, education, agriculture and recommendations. It emphasizes the importance of user involvement, personalization and domain expertise in developing AI systems that non-experts can understand and trust.
Poster: Perspectives on Increasing Competency in Using Digital Practices and ...Katja Reuter, PhD
We believe that the quality and efficiency of all phases of the clinical and translational research (CTR) process can potentially be increased by using digital practices and tools in open and networked contexts. However, most CT researchers lack the training to take advantage of the benefits that the Internet and the social Web provide. Standardized training in digital practices and tools (Digital Scholarship) to conduct CTR has not been formalized through structured curriculum, learning approaches, and evaluation. Our overall goal is to develop a robust curriculum to train CTR researchers in digital scholarship. Here we present preliminary data from a qualitative study that describes the range of key stakeholders’ perspectives on the need to: (A) formalize educational efforts in digital scholarship among CTR trainees; and (B) develop an educational framework that defines core competencies, methods, and evaluation methods. Presented at Translational Science 2018 conference in Washington, DC on April 20, 2018.
THE USE OF CLOUD COMPUTING SYSTEMS IN HIGHER EDUCATION; The Lived Experiences of Faculty
Dr. Joseph K. Adjei
School of Technology (SOT)
Ghana Institute of Management and Public Administration (GIMPA)
2nd International Conference of the African Virtual University
Seffah iess11 keynote the human side of service scienceIESS
This keynote talk discusses the need for a human-centric quality model for services and service systems. It motivates moving beyond traditional software and service quality models to account for new factors like trust, privacy, and universality. The talk proposes a quality model with three components: factors, criteria to measure sub-factors, and qualitative/quantitative measures. It suggests using personas, scenarios, and design patterns as tools to model user experiences and derive service systems. The quality model can then assess design artifacts and final services/systems to predict quality in use. The talk aims to bridge technical and human aspects of service design.
This document discusses personas and how they are used as a design tool in interaction design. It provides information on what personas are, how they are developed based on research, their typical elements, and different types of personas. Examples of personas are also provided at the end to illustrate how personas describe fictional people based on research in order to represent different user groups and their goals. The document aims to explain the purpose and process of developing personas to inform interaction design.
Leveraging Graph Neural Networks for User Profiling: Recent Advances and Open...Erasmo Purificato
Slide of the tutorial entitled "Leveraging Graph Neural Networks for User Profiling: Recent Advances and Open Challenges" held at CIKM'23: 32nd ACM International Conference on Information and Knowledge Management (October 21, 2023 | Birmingham, United Kingdom)
This document summarizes a presentation given by Katrien Verbert on interactive recommender systems. It provides an overview of Verbert's research group at KU Leuven, which focuses on recommender systems, visualization, and intelligent user interfaces. The presentation describes various techniques for building interactive recommender systems, including explaining recommendations to users, enabling user interaction with the recommendation process, and addressing challenges like diversity, cold starts, and context awareness. It also summarizes several studies conducted by Verbert and collaborators on interactive music and research talk recommender systems.
We the humans are surrounded with immense unprecedented wealth of information which are available as documents, database or other resources. The access to this information is difficult as by having the information it is not necessary that it could be searched or extracted by the activity we are using. The search engines available should be also customized to handle such queries, sometime the search engines are also not aware of the information they have within the system. The method known as keyword extraction and clustering is introduced which answers this shortcoming by spontaneously recommending documents that are related to users’ current activities. When the communication takes place the important text can be extracted from the conversation and the words extracted are grouped and then are matched with the parts in the document. This method uses Natural Language Processing for extracting of keywords and making the subgroup that is a meaningful statement from the group, another method used is the Hierarchical Clustering for creating clusters form the keywords, here the similarity of two keywords is measured using the Euclidean distance. This paper reviews the various methods for the system.
Redesigning the Open Access Institutional RepositoryEdward Luca
This lecture presents a redesign project of UTS's institutional repository, OPUS. It explains some of the challenges faced by libraries in ensuring eRepository participation, and investigates three user groups - academics, librarians, and information seekers. User experience principles are used to address issues around navigation, terminology, and visual identity.
Presented as a guest lecture to Designing for the Web (Spring 2016) students.
Interactive recommender systems: opening up the “black box”Katrien Verbert
This document summarizes a presentation given by Katrien Verbert on interactive recommender systems. It discusses how recommender systems are typically "black boxes" that do not explain their recommendations to users. The presentation aims to open up this black box by exploring ways to increase transparency, user control, and interaction with recommender systems. Examples of interactive recommender systems that allow users to explore the recommendation process and provide explanations are described. Research on developing and evaluating such interactive systems through multiple user studies is summarized. The objective is to enhance user trust and engagement with recommender systems.
This document discusses recommender systems and approaches used at Netflix. It covers collaborative filtering using user-user and item-item methods, content-based recommendations using item attributes, and hybrid approaches. It provides examples of how Netflix uses collaborative filtering to generate personalized genre rows and social recommendations. Netflix combines many data sources and machine learning models to power its highly personalized recommendation engine.
This document summarizes Katrien Verbert's presentation on interactive recommender systems. The presentation covered several topics:
1) Different types of recommendation techniques including collaborative filtering, content-based filtering, and knowledge-based filtering.
2) Research on interactive recommender systems that aim to increase transparency, user control, and diversity of recommendations.
3) Several user studies conducted on interactive recommender systems that explored talks and conferences, finding that explanations and various levels of user control can impact user experience.
This document summarizes quality criteria that have been proposed for mixed methods research. It discusses quality standards for quantitative and qualitative research separately, and notes the lack of agreement on criteria for qualitative research. The document then examines several proposed sets of criteria for mixed methods research, focusing on appropriate research design, rigorous data collection and analysis, and thorough integration of quantitative and qualitative components. It concludes by describing two studies that sought views from social scientists on quality evaluation in mixed methods studies.
This thesis proposes designing and developing a personalized country recommender system. It outlines introducing the problem motivation and research questions. The document then reviews the state of the art on recommender systems including definitions, data sources, approaches (collaborative filtering, content-based filtering, hybrid filtering), and evaluation metrics. It describes the methodology which includes collecting a training dataset, implementing recommender algorithms (SVD, KNN, co-clustering), and system design. The results and evaluation of the system are then presented.
This thesis examines the digital competence of pre-service and in-service teachers in Anhui Province, China. It includes 4 chapters that 1) review relevant frameworks and literature, 2) assess teachers' digital competence using a validated questionnaire, 3) analyze the results, and 4) will formulate proposals for improving teachers' digital skills. The assessment found teachers' digital competence was at a medium level, with some differences between pre-service and in-service groups. Factors like age, education level, and experience influenced skills. Overall, the study evaluated teachers' abilities and will make recommendations to enhance digital pedagogy in Anhui.
This document provides an overview of personas in libraries and how they can be used. It discusses creating personas for different user groups like public library patrons, students, faculty, and internal library staff. The document also covers how to develop personas through qualitative research methods like interviews and surveys, and how to validate personas with quantitative data from website analytics and surveys. Personas can then be used to help make strategic decisions, design services, and test usability within libraries.
Contextual model of recommending resources on an academic networking portalcsandit
Artificial Intelligence techniques have been instrumental in helping users to handle the large
amount of information on the Internet. The idea of recommendation systems, custom search
engines, and intelligent software has been widely accepted among users who seek assistance in
searching, sorting, classifying, filtering and sharing this vast quantity of information. In this
paper, we present a contextual model of recommendation engine which keeping in mind the
context and activities of a user, recommends resources in an academic networking portal. The
proposed method uses the implicit method of feedback and the concepts relationship hierarchy
to determine the similarity between a user and the resources in the portal. The proposed
algorithm has been tested on an academic networking portal and the results are convincing.
CONTEXTUAL MODEL OF RECOMMENDING RESOURCES ON AN ACADEMIC NETWORKING PORTALcscpconf
Artificial Intelligence techniques have been instrumental in helping users to handle the large amount of information on the Internet. The idea of recommendation systems, custom search engines, and intelligent software has been widely accepted among users who seek assistance insearching, sorting, classifying, filtering and sharing this vast quantity of information. In thispaper, we present a contextual model of recommendation engine which keeping in mind the context and activities of a user, recommends resources in an academic networking portal. Theproposed method uses the implicit method of feedback and the concepts relationship hierarchy to determine the similarity between a user and the resources in the portal. The proposed algorithm has been tested on an academic networking portal and the results are convincing
A Literature Review of Quantitative Persona CreationMinjoon Kim
This paper provides a literature review of quantitative persona creation (QPC), which uses algorithmic methods to create personas from numerical and textual data. The paper analyzes 47 research papers on QPC. It identifies three stages of evolution for QPC methods, from using basic clustering techniques on survey data to leveraging large datasets from social media and APIs. Key trends include higher automation, interactive persona systems, and combining automatic and manual methods. The paper also discusses research gaps in developing standards, addressing ethics concerns, and retaining benefits of qualitative persona creation. It concludes with recommendations to advance the field of QPC.
Human-centered AI: how can we support lay users to understand AI?Katrien Verbert
The document summarizes research on human-centered AI and how to support lay users in understanding AI. It discusses various research projects that aim to explain model outcomes to increase user trust and acceptance. It explores how personal characteristics like need for cognition can impact the effectiveness of explanations. The research also looks at different application domains for AI like healthcare, education, agriculture and recommendations. It emphasizes the importance of user involvement, personalization and domain expertise in developing AI systems that non-experts can understand and trust.
Poster: Perspectives on Increasing Competency in Using Digital Practices and ...Katja Reuter, PhD
We believe that the quality and efficiency of all phases of the clinical and translational research (CTR) process can potentially be increased by using digital practices and tools in open and networked contexts. However, most CT researchers lack the training to take advantage of the benefits that the Internet and the social Web provide. Standardized training in digital practices and tools (Digital Scholarship) to conduct CTR has not been formalized through structured curriculum, learning approaches, and evaluation. Our overall goal is to develop a robust curriculum to train CTR researchers in digital scholarship. Here we present preliminary data from a qualitative study that describes the range of key stakeholders’ perspectives on the need to: (A) formalize educational efforts in digital scholarship among CTR trainees; and (B) develop an educational framework that defines core competencies, methods, and evaluation methods. Presented at Translational Science 2018 conference in Washington, DC on April 20, 2018.
THE USE OF CLOUD COMPUTING SYSTEMS IN HIGHER EDUCATION; The Lived Experiences of Faculty
Dr. Joseph K. Adjei
School of Technology (SOT)
Ghana Institute of Management and Public Administration (GIMPA)
2nd International Conference of the African Virtual University
THE USE OF CLOUD COMPUTING SYSTEMS IN HIGHER EDUCATION; The Lived Experien...
Umap15fernandeztobias slides
1. On the use of
cross-domain user preferences and
personality traits
in collaborative filtering
Ignacio Fernández-Tobías, Iván Cantador
{ignacio.fernandezt, ivan.cantador}@uam.es
Information Retrieval Group
Universidad Autónoma de Madrid, Spain
2. 1
On the use of cross-domain user preferences and personality traits in collaborative filtering
23rd Conference on User Modelling, Adaptation and Personalization (UMAP 2013)
3rd July 2015, Dublin, Ireland
The Five-Factor model of personality
Openness (OPE)
Conscientiousness (COS)
Extraversion (EXT)Agreeableness (AGR)
Neuroticism (NEU)
cautious/consistent vs.
curious/inventive
careless/easy-going vs.
organized/efficient
solitary/reserved vs.
outgoing/energetic
cold/unkind vs.
friendly/compassionate
secure/calm vs.
unconfident/nervous
tendency to intellectual
curiosity, creativity and
preference for novelty and
variety of experiences
tendency to show self-discipline and
aim for personal achievements, and to
have an organized (not spontaneous)
and dependable behavior
tendency to seek stimulation in the
company of others, and to put energy
in finding positive emotions
tendency to be kind,
concerned, truthful and
cooperative towards others
tendency to experience
unpleasant emotions, and
low degree of emotional
stability and impulse control
3. 2
On the use of cross-domain user preferences and personality traits in collaborative filtering
23rd Conference on User Modelling, Adaptation and Personalization (UMAP 2013)
3rd July 2015, Dublin, Ireland
• Music
• (Rentfrow & Gosling , 2003): reflective people (high OPE) jazz, blues, classical music;
energetic people (high EXT and high AGR) rap, hip-hop, funk, electronic music
• (Rawlings & Ciancarelli, 1997): high OPE high diversity of music preferences;
high EXT popular music
• Movies and TV shows
• (Chausson, 2010): high OPE comedy & fantasy movies; high COS action movies;
high NEU romantic movies
• (Odić et al., 2013): emotional patterns induced by movies as functions of EXT, AGR, NEU
• Multiple domains (music, movies-TV shows, books-magazines, ...)
• (Rentfrow et al., 2011): relations between preferences and personality-based
categories, e.g. aesthetic, cerebral, communal, dark, and thrilling
• (Kosinski et al., 2012): relations between preferences and personality for certain
websites and website categories
• (Cantador et al., 2013): association rules between preferences and personality factors
User preferences and personality traits
4. 3
On the use of cross-domain user preferences and personality traits in collaborative filtering
23rd Conference on User Modelling, Adaptation and Personalization (UMAP 2013)
3rd July 2015, Dublin, Ireland
Relations between user preferences and
personality traits may be exploited in
personalization and recommendation services
5. 4
On the use of cross-domain user preferences and personality traits in collaborative filtering
23rd Conference on User Modelling, Adaptation and Personalization (UMAP 2013)
3rd July 2015, Dublin, Ireland
• Addressing cold-start situations
• (Hu & Pu, 2010; Tkalčič et al., 2011): user similarities in CF with personality information
• Mitigating the sparsity problem
• (Hu & Pu, 2011): increasing the density of rating matrices by means of personality data
• Facilitating the user preference elicitation
• (Elahi et al., 2013): exploiting the user’s personality to identify the items to rate
• Improving recommendation accuracy
• (Nunes et al., 2009): user models composed of personality factors & facets
• (Roshchina , 2012): personality-aware CB and CF recommendation models
• (Fernández-Tobías & Cantador, 2014): incorporating both user preferences and
personality factors into CF heuristics
• (Wu & Chen, 2015): integrating implicitly acquired personality profiles into CF
heuristics
•
Exploiting user personality in recsys
6. 5
On the use of cross-domain user preferences and personality traits in collaborative filtering
23rd Conference on User Modelling, Adaptation and Personalization (UMAP 2013)
3rd July 2015, Dublin, Ireland
• Addressing cold-start situations
• (Hu & Pu, 2010; Tkalčič et al., 2011): user similarities in CF with personality information
• Mitigating the sparsity problem
• (Hu & Pu, 2011): increasing the density of rating matrices by means of personality data
• Facilitating the user preference elicitation
• (Elahi et al., 2013): exploiting the user’s personality to identify the items to rate
• Improving recommendation accuracy
• (Nunes et al., 2009): user models composed of personality factors & facets
• (Roshchina , 2012): personality-aware CB and CF recommendation models
• (Fernández-Tobías & Cantador, 2014): incorporating both user preferences and
personality factors into CF heuristics
• (Wu & Chen, 2015): integrating implicitly acquired personality profiles into CF
heuristics
• Enhancing cross-domain recommendations
Exploiting user personality in recsys
7. 6
On the use of cross-domain user preferences and personality traits in collaborative filtering
23rd Conference on User Modelling, Adaptation and Personalization (UMAP 2013)
3rd July 2015, Dublin, Ireland
Cross-domain recommenders systems
aim to generate or enhance recommendations in a target
domain by exploiting knowledge from source domains
Cantador, I., Fernández-Tobías, I., Berkovsky, S., Cremonesi, P. 2015. Cross-domain
Recommender Systems. In Ricci, F., Rokach, L., Shapira, B., Kantor, P. B. (Eds.), Recommender
Systems Handbook - 2nd edition. To appear
Target
domain
Source
domain
+
knowledge
aggregation
target domain
recommendations
Target
domain
Source
domain
knowledge
linkage/transfer
target domain
recommendations
Aggregating knowledge Linking/transferring knowledge
8. 7
On the use of cross-domain user preferences and personality traits in collaborative filtering
23rd Conference on User Modelling, Adaptation and Personalization (UMAP 2013)
3rd July 2015, Dublin, Ireland
How user preferences in a source domain and
personality factors can be exploited to provide
effective recommendations in a target domain?
9. 8
On the use of cross-domain user preferences and personality traits in collaborative filtering
23rd Conference on User Modelling, Adaptation and Personalization (UMAP 2013)
3rd July 2015, Dublin, Ireland
How user preferences in a source domain and
personality factors can be exploited to provide
effective recommendations in a target domain?
1. Personality-based CF heuristics
2. Personality-based CF factor models
10. 9
On the use of cross-domain user preferences and personality traits in collaborative filtering
23rd Conference on User Modelling, Adaptation and Personalization (UMAP 2013)
3rd July 2015, Dublin, Ireland
Proposed personality-based CF heuristics
rating estimation
hybrid user similarity
personality-based
COS
PEA
SPE
preference-based
CF
11. 10
On the use of cross-domain user preferences and personality traits in collaborative filtering
23rd Conference on User Modelling, Adaptation and Personalization (UMAP 2013)
3rd July 2015, Dublin, Ireland
• Matrix factorization (MF) recommendation
• Cross-domain MF recommendation
• Personality-based cross-domain MF recommendation
Proposed personality-based CF factor models
12. 11
On the use of cross-domain user preferences and personality traits in collaborative filtering
23rd Conference on User Modelling, Adaptation and Personalization (UMAP 2013)
3rd July 2015, Dublin, Ireland
• A dataset obtained from the MyPersonality project*
• Facebook likes for movies, music items, and books
• Five-factor personality profiles from psychometric questionnaires
‐ Revised NEO Personality Inventory (NEO PI-R): 60-240 questions
• Statistics (updated from those reported in the paper)
Experiments - dataset
* http://mypersonality.org. We thank David Stillwell & Michal Kosinski for their support
13. 12
On the use of cross-domain user preferences and personality traits in collaborative filtering
23rd Conference on User Modelling, Adaptation and Personalization (UMAP 2013)
3rd July 2015, Dublin, Ireland
• Cross-domain situations
• user overlap
• no user overlap
• Cold-start situations in the target domain
• extreme: 0 training ratings per user
• moderate: from 1 to 10 training ratings per user
• Additional user information
• source-domain ratings
• personality factors
Experiments - evaluation scenarios
14. 13
On the use of cross-domain user preferences and personality traits in collaborative filtering
23rd Conference on User Modelling, Adaptation and Personalization (UMAP 2013)
3rd July 2015, Dublin, Ireland
• Baseline recommendation methods
• Random
• Most popular
• Item-based kNN (k=∞) and User-based kNN (k=50, but others tested as well)
• MF (10 factors, but others tested as well)
• Proposed recommendation methods
• Personality-based heuristics (λ=0, 0.1, …, 0.9; note that λ=1 is user-based
kNN)
• Personality-based MF models
• Recommendation performance metrics
• Accuracy and ranking metrics for positive feedback: mainly MAP, but other
metrics computed (Half-life utility, Mean Percentage Ranking)
• Coverage and novelty
• 5-fold cross validation + statistical significance tests
Experiments - evaluation setting
15. 14
On the use of cross-domain user preferences and personality traits in collaborative filtering
23rd Conference on User Modelling, Adaptation and Personalization (UMAP 2013)
3rd July 2015, Dublin, Ireland
• In “extreme” cold-start situations
• Recommendation methods
‐ MF is not capable of providing recommendations
‐ Personality-based MF outperforms Popularity
• Cross domains:
‐ Movies Books
‐ Music + personality Movies
‐ Movies + personality Music
• In “moderate” cold-start situations
• Recommendation methods:
‐ Both MF and personality-based MF clearly outperform Popularity
‐ Personality-based MF performs slightly better than MF
• Cross domains:
‐ Music + personality Movies (small improvements)
‐ Movies + personality Music (clear improvements)
Experiments - result highlights
16. 15
On the use of cross-domain user preferences and personality traits in collaborative filtering
23rd Conference on User Modelling, Adaptation and Personalization (UMAP 2013)
3rd July 2015, Dublin, Ireland
• Both user personality factors and cross-domain preferences
allow dealing with cold-start situations
• best results achieved when crossing the movies and music domains
• experiments conducted with likes instead of numeric ratings
• Reasonable recommendation performance improvements
• moderate better accuracy- and ranking-based performance
• better diversity and coverage performance
• Difficulty to represent user personality in an effective way
(for recommendation purposes)
• discretization of personality factor values
Conclusions
17. 16
On the use of cross-domain user preferences and personality traits in collaborative filtering
23rd Conference on User Modelling, Adaptation and Personalization (UMAP 2013)
3rd July 2015, Dublin, Ireland
• Personality factors
• (Hu & Pu, 2010; Tkalčič et al., 2011; Fernández-Tobías & Cantador, 2014): user
profiles composed of numeric values of personality factors
• Personality facets
• (Nunes et al., 2009): e.g. OPE facets: imagination, artistic interests,
emotionality, ...
• Personality categories
• (Rentfrow & Gosling , 2003): e.g. reflective and energetic people (user level)
• (Rentfrow et al., 2011): e.g. aesthetic, cerebral and thrilling contents (item
level)
• Personality user stereotypes
• (Lin & McLeod, 2002): human temperaments from Keirsey’s theory: guardian,
idealist, rational, artisan
Future work - Personality user models
18. 17
On the use of cross-domain user preferences and personality traits in collaborative filtering
23rd Conference on User Modelling, Adaptation and Personalization (UMAP 2013)
3rd July 2015, Dublin, Ireland
Future work - User attributes
• age and gender
• (Fernández-Tobías & Cantador, 2014): different correlations between
preferences and personality profiles depending on the users’ gender and age
- Erikson’s psychosocial stages (1950)
• educational attainment
• e.g., “people with high levels of education may be more open-minded, and
thus have larger and more diverse sets of preferences”
• others…
19. On the use of
cross-domain user preferences and
personality traits
in collaborative filtering
Ignacio Fernández-Tobías, Iván Cantador
{ignacio.fernandezt, ivan.cantador}@uam.es
Information Retrieval Group
Universidad Autónoma de Madrid, Spain