Giovanni Semeraro - Profile on Academia.edu (original) (raw)

Papers by Giovanni Semeraro

User Modeling and User-Adapted Interaction

Users of online recipe websites tend to prefer unhealthy foods. Their popularity undermines the h... more Users of online recipe websites tend to prefer unhealthy foods. Their popularity undermines the healthiness of traditional food recommender systems, as many users lack nutritional knowledge to make informed food decisions. Moreover, the presented information is often unrelated to nutrition or difficult to understand. To alleviate this, we present a methodology to generate natural language justifications that emphasize the nutritional content, health risks, or benefits of recommended recipes. Our framework takes a user and two recipes as input and produces an automatically generated natural language justification as output, based on the user’s characteristics and the recipes’ features, following a knowledge-based recommendation approach. We evaluated our methodology in two crowdsourcing studies. In Study 1 ($$N=502$$ N = 502 ), we compared user food choices for two personalized recommendation approaches, based on either a (1) single-style justification or (2) comparative justificatio...

Proceedings of the 2012 International Conference on Semantic Technologies Meet Recommender Systems & Big Data - Volume 919

User Modeling and User-Adapted Interaction

This paper introduces a methodology to generate review-based natural language justifications supp... more This paper introduces a methodology to generate review-based natural language justifications supporting personalized suggestions returned by a recommender system. The hallmark of our strategy lies in the fact that natural language justifications are adapted to the different contextual situations in which the items will be consumed. In particular, our strategy relies on the following intuition: Just like the selection of the most suitable item is influenced by the contexts of usage, a justification that supports a recommendation should vary as well. As an example, depending on whether a person is going out with her friends or her family, a justification that supports a restaurant recommendation should include different concepts and aspects. Accordingly, we designed a pipeline based on distributional semantics models to generate a vector space representation of each context. Such a representation, which relies on a term-context matrix, is used to identify the most suitable review exce...

HELENA: An intelligent digital assistant based on a Lifelong Health User Model

Information Processing & Management

Adjunct Proceedings of the 30th ACM Conference on User Modeling, Adaptation and Personalization

The phenomenon of hate messages on the web is unfortunately in continuous expansion and evolution... more The phenomenon of hate messages on the web is unfortunately in continuous expansion and evolution. Even if the big companies that offer their users a social network service have expressly included in their terms of services rules against hate messages, they are still produced at a huge rate. Therefore, moderators are often employed to monitor these platforms and use their critical skills to decide if the content is offensive or not. Unfortunately, this censorship process is complex and costly in terms of human resources. The system we propose in this work is a system that supports moderators by providing them a set of candidate elements to censor with annexed explanations in natural language. It will then be a task of the human operator to understand if to proceed with the censorship and eventually supply feedback to the result of the classification algorithm to extend its data set of examples and improve its future performances. The proposed system has been designed to merge information coming from data, syntactic tags and a manually annotated lexicon. The messages are then processed through deep learning approaches based on both transformer and deep neural network architecture. The output is consequently supported by an explanation in a human-like form. The model has been evaluated on three state-of-the-art datasets showing excellent effectiveness and clear and understandable explanations.

Document Indexing based on Word Sense Disambiguation, WordNet and Wikipedia

AI*IA 2008, 10° Convegno dell'Associazione Italiana per l'Intelligenza Artificiale, 2008

Humanoid Robots and Conversational Recommender Systems: a Preliminary Study

2020 IEEE Conference on Evolving and Adaptive Intelligent Systems (EAIS), 2020

Conversational Recommender Systems (CoRSs) implement a paradigm in which users can interact with ... more Conversational Recommender Systems (CoRSs) implement a paradigm in which users can interact with the system to define their preferences and discover items that best fit their needs. When the CoRS is implemented as a dialog agent, user and recommender interact by exchanging text messages. However, there is little evidence on how effective the interaction is when the CoRS is implemented through a Social Humanoid Robot. In this paper, we evaluate the possibility of introducing an interface based on a Social Humanoid Robot in ConveRSE, a domain-independent framework for the development of Conversational Recommender Systems. The novel interface will be compared against the existing chatbot-based one. The objective is to discover whether the framework can adapt to the new interface without worsening user experience and accuracy. We carried out a preliminary study, which involved 20 subjects. Results proved that, even though there are differences in how users approach the system using the two interfaces, there is no significant difference in its performance.

In this work we propose Ask Me Any Rating (AMAR), a novel content-based recommender system based ... more In this work we propose Ask Me Any Rating (AMAR), a novel content-based recommender system based on deep neural networks which is able to produce top-N recommendations leveraging user and item embeddings which are learnt from textual information describing the items. A comprehensive experimental evaluation conducted on stateof-the-art datasets showed a significant improvement over all the baselines taken into account.

Constraint-based recommenders support customers in identifying relevant items from complex item a... more Constraint-based recommenders support customers in identifying relevant items from complex item assortments. In this paper we present WeeVis, a constraint-based environment that can be applied in different scenarios in the e-government domain. WeeVis supports collaborative knowledge acquisition for recommender applications in a MediaWiki-based context. This paper shows how Wiki pages can be extended with recommender applications and how the environment uses intelligent mechanisms to support users in identifying the optimal solutions to their needs. An evaluation shows a performance overview with different knowledge bases.

Do You Feel Blue? Detection of Negative Feeling from Social Media

AI*IA 2017 Advances in Artificial Intelligence, 2017

The blue feeling is the sensation which affects people when they feel down, depressed, sad and mo... more The blue feeling is the sensation which affects people when they feel down, depressed, sad and more generally when they are in a bad feeling state. In some cases, it is a recurring situation in their everyday life and it can be the first symptom of more complex psychological diseases such as depression. In the last decade, as consequence of the quick increase of detected cases of depression in children and teenagers, it has become very important to find strategies for a timely detection of this pathology. In this work, we describe a model that can support the detection task, by identifying some warning scenarios of blue feeling. The proposed architecture is composed by modules focused on different aspects that characterize the scenario: changes in heart rate, reduction of sleep, reduction of activities performed, increases of use of negative phrases and words. In particular, in this paper, we describe the approach adopted to analyze users posts on social media networks (SMNs) by using natural language processing techniques. The proposed approach is evaluated through an experimental session over a dataset of Facebook posts. The results show good performance in the detection of negative feeling.

Proceedings of the Joint Workshop on Interfaces and Human Decision Making for Recommender Systems, IntRS 2015, Vienna, Austria, September 19, 2015

Proceedings of the 5th Joint Workshop on Interfaces and Human Decision Making for Recommender Systems, IntRS 2018, co-located with ACM Conference on Recommender Systems (RecSys 2018), Vancouver, Canada, October 7, 2018

The popularity of social robots is steadily increasing, mainly due to the interesting impact they... more The popularity of social robots is steadily increasing, mainly due to the interesting impact they have in several application domains. In this paper, we propose the use of Pepper Robot as an interface of a recommender system for tourism. In particular, we used the robot to interact with the users and to provide them with personalized recommendations about hotels, restaurants, and points of interest in the area. The personalization mechanism encoded in the social robot relies on soft biometrics traits automatically recognized by the robot, as age and gender, user interests and personal facets. All these data are used to feed a neural network that returns as output the most suitable recommendations for the target user. To evaluate the effectiveness of the interaction driven by a social robot, we carried out a user study whose goal was to evaluate: (1) how the robot affects the perceived accuracy of the recommendations; (2) how the user experience and the engagement vary by interacting...

Interfaces and Human Decision Making for Recommender Systems

Fourteenth ACM Conference on Recommender Systems, 2020

As an interactive intelligent system, recommender systems are developed to give recommendations t... more As an interactive intelligent system, recommender systems are developed to give recommendations that match users’ preferences. Since the emergence of recommender systems, a large majority of research focuses on objective accuracy criteria and less attention has been paid to how users interact with the system and the efficacy of interface designs from users’ perspectives. The field has reached a point where it is ready to look beyond algorithms, into users’ interactions, decision making processes, and overall experience. The series of workshops on Interfaces and Human Decision Making for Recommender Systems focuses on the ”human side” of recommender systems. The goal of the research stream featured at the workshop is to improve users’ overall experience with recommender systems by integrating different theories of human decision making into the construction of recommender systems and exploring better interfaces for recommender systems. In this summary, we introduce 7th Joint Workshop on Interfaces and Human Decision Making for Recommender Systems at RecSys’20, review its history, and discuss most important topics considered at the workshop.

Towards Emotion-aware Recommender Systems: an Affective Coherence Model based on Emotion-driven Behaviors

Expert Systems with Applications, 2021

Abstract Decision making is the cognitive process of identifying and choosing alternatives based ... more Abstract Decision making is the cognitive process of identifying and choosing alternatives based on preferences, beliefs, and degree of importance given by the decision maker to objects or actions. For instance, choosing which movie to watch is a simple, small-sized decision-making process. Recommender systems help people to make this kind of choices, usually by computing a short list of suggestions that reduces the space of possible options. These systems are strongly based on the knowledge of user preferences but, in order to fully support people, they should be grounded on a holistic view of the user behavior, that includes also how emotions, mood, and personality traits influence her choosing patterns. In this work, we investigate how to include emotional aspects in the recommendation process. We suggest that the affective state of the user, defined by a set of emotions (e.g., joy, surprise), constitutes part of choosing situation that should be taken into account when modeling user preferences. The main contribution of the paper is a general emotion-aware computational model based on affective user profiles in which each preference, such as a 5-star rating on a movie, is associated with the affective state felt by the user at the time when that preference was collected. The model estimates whether an unseen item is suitable for the current affective state of the user, by computing an affective coherence score that takes into account both the affective user profile and not-affective item features. The approach has been implemented into an Emotion-aware Music Recommender System, whose effectiveness has been assessed by performing in-vitro experiments on two benchmark datasets. The main outcome is that our system showed improved accuracy of recommendations compared to baselines which include no affective information in the recommendation model.

IEEE Access, 2020

Recommender systems (RSs) are systems that produce individualized recommendations as output or dr... more Recommender systems (RSs) are systems that produce individualized recommendations as output or drive the user in a personalized way to interesting or useful objects in a space of possible options. Recently, RSs emerged as an effective support for decision making. However, when people make decisions, they usually take into account different and often conflicting information such as preferences, long-term goals, context, and their current condition. This complexity is often ignored by RSs. In order to provide an effective decision-making support, a RS should be ''holistic'', i.e., it should rely on a complete representation of the user, encoding heterogeneous user features (such as personal interests, psychological traits, health data, social connections) that may come from multiple data sources. However, to obtain such holistic recommendations some steps are necessary: first, we need to identify the goal of the decision-making process; then, we have to exploit common-sense and domain knowledge to provide the user with the most suitable suggestions that best fit the recommendation scenario. In this article, we present a methodological framework that can drive researchers and developers during the design process of this kind of ''holistic'' RS. We also provide evidence of the framework validity by presenting the design process and the evaluation of a food RS based on holistic principles.

Information Systems, 2017

HealthNet (HN) is a social network that brings together patients with similar health conditions. ... more HealthNet (HN) is a social network that brings together patients with similar health conditions. HN helps users in finding a solution to their health problems by suggesting doctors and health facilities that best fit the patient profile. Indeed, the core component of HN is a recommender system that suggests patients similar to the target user and supports the choice of the doctor and the hospital for a specific condition. The recommendation algorithm first computes similarities among patients, and then generates a ranked list of doctors and hospitals for a given patient profile by exploiting health data shared by the community. The HN typical user can find the most similar patients, can look how they treated their diseases, and can receive suggestions for solving her condition. In order to facilitate the interaction with the system and improve the recommendation step, the patient can express her health status by a natural-language sentence. The system analyzes the sentence and identifies the most relevant medical area (e.g., orthopedics, neurology, allergology, etc.) for that specific case, and uses this information for the recommendation task. Currently HN is in alpha version and only for Italian users, but in the future we want to extend the platform to other languages. We carried out both an in-vitro experimental evaluation to assess the effectiveness of the module for analyzing natural language descriptions provided by users as well as the recommender system to suggest the right

Information Sciences, 2017

Cross-lingual data linking is the problem of establishing links between resources, such as places... more Cross-lingual data linking is the problem of establishing links between resources, such as places, services, or movies, which are described in different languages. In cross-lingual data linking it is often the case that very short descriptions have to be matched, which makes the problem even more challenging. This work presents a method named TRanslation-based Explicit Semantic Analysis (tr-esa) to represent and match short textual descriptions available in different languages. tr-esa translates short descriptions in any given language into a pivot language by exploiting a machine translation tool. Then, it generates a Wikipedia-based representation of the translated text by using the Explicit Semantic Analysis technique. The resulting representations are used to match short descriptions in different languages. The method is incorporated in CroSeR (Cross-lingual Service Retrieval), an interactive data linking tool that recommends potential matches to users. We compared results coming from an in-vitro evaluation on a gold standard consisting of five datasets in different languages, with an in-vivo experiment that involved human experts supported by CroSeR. The in-vivo evaluation confirmed the results of the in-vitro evaluation and the overall effectiveness of the proposed method.

Emotion Detection Techniques for the Evaluation of Serendipitous Recommendations

Human–Computer Interaction Series, 2016

Recommender systems analyze a user’s past behavior, build a user profile that stores information ... more Recommender systems analyze a user’s past behavior, build a user profile that stores information about her interests, maybe find others who have a similar profile, and use that information to find potentially interesting items. The main limitation of this approach is that provided recommendations are accurate, because they match the user profile, but not useful as they fall within the existing range of user interests. This drawback is known as overspecialization. New methods are being developed to compute serendipitous recommendations, i.e. unexpected suggestions that stimulate the user curiosity toward potentially interesting items she might not have otherwise discovered. The evaluation of those methods is not simple: there is a level of emotional response associated with serendipitous recommendations that is difficult to measure. In this chapter, we discuss the role of emotions in recommender systems research, with focus on their exploitation as implicit feedback on suggested items. Furthermore, we describe a user study which assesses both the acceptance and the perception of serendipitous recommendations, through the administration of questionnaires and the analysis of users’ emotions. Facial expressions of users receiving recommendations are analyzed to evaluate whether they convey a mixture of emotions that helps to measure the perception of serendipity of recommendations. The results showed that positive emotions such as happiness and surprise are associated with serendipitous suggestions.

Proceedings of the Second Italian Conference on Computational Linguistics CLiC-it 2015

User Modeling and User-Adapted Interaction

Users of online recipe websites tend to prefer unhealthy foods. Their popularity undermines the h... more Users of online recipe websites tend to prefer unhealthy foods. Their popularity undermines the healthiness of traditional food recommender systems, as many users lack nutritional knowledge to make informed food decisions. Moreover, the presented information is often unrelated to nutrition or difficult to understand. To alleviate this, we present a methodology to generate natural language justifications that emphasize the nutritional content, health risks, or benefits of recommended recipes. Our framework takes a user and two recipes as input and produces an automatically generated natural language justification as output, based on the user’s characteristics and the recipes’ features, following a knowledge-based recommendation approach. We evaluated our methodology in two crowdsourcing studies. In Study 1 ($$N=502$$ N = 502 ), we compared user food choices for two personalized recommendation approaches, based on either a (1) single-style justification or (2) comparative justificatio...

Proceedings of the 2012 International Conference on Semantic Technologies Meet Recommender Systems & Big Data - Volume 919

User Modeling and User-Adapted Interaction

This paper introduces a methodology to generate review-based natural language justifications supp... more This paper introduces a methodology to generate review-based natural language justifications supporting personalized suggestions returned by a recommender system. The hallmark of our strategy lies in the fact that natural language justifications are adapted to the different contextual situations in which the items will be consumed. In particular, our strategy relies on the following intuition: Just like the selection of the most suitable item is influenced by the contexts of usage, a justification that supports a recommendation should vary as well. As an example, depending on whether a person is going out with her friends or her family, a justification that supports a restaurant recommendation should include different concepts and aspects. Accordingly, we designed a pipeline based on distributional semantics models to generate a vector space representation of each context. Such a representation, which relies on a term-context matrix, is used to identify the most suitable review exce...

HELENA: An intelligent digital assistant based on a Lifelong Health User Model

Information Processing & Management

Adjunct Proceedings of the 30th ACM Conference on User Modeling, Adaptation and Personalization

The phenomenon of hate messages on the web is unfortunately in continuous expansion and evolution... more The phenomenon of hate messages on the web is unfortunately in continuous expansion and evolution. Even if the big companies that offer their users a social network service have expressly included in their terms of services rules against hate messages, they are still produced at a huge rate. Therefore, moderators are often employed to monitor these platforms and use their critical skills to decide if the content is offensive or not. Unfortunately, this censorship process is complex and costly in terms of human resources. The system we propose in this work is a system that supports moderators by providing them a set of candidate elements to censor with annexed explanations in natural language. It will then be a task of the human operator to understand if to proceed with the censorship and eventually supply feedback to the result of the classification algorithm to extend its data set of examples and improve its future performances. The proposed system has been designed to merge information coming from data, syntactic tags and a manually annotated lexicon. The messages are then processed through deep learning approaches based on both transformer and deep neural network architecture. The output is consequently supported by an explanation in a human-like form. The model has been evaluated on three state-of-the-art datasets showing excellent effectiveness and clear and understandable explanations.

Document Indexing based on Word Sense Disambiguation, WordNet and Wikipedia

AI*IA 2008, 10° Convegno dell'Associazione Italiana per l'Intelligenza Artificiale, 2008

Humanoid Robots and Conversational Recommender Systems: a Preliminary Study

2020 IEEE Conference on Evolving and Adaptive Intelligent Systems (EAIS), 2020

Conversational Recommender Systems (CoRSs) implement a paradigm in which users can interact with ... more Conversational Recommender Systems (CoRSs) implement a paradigm in which users can interact with the system to define their preferences and discover items that best fit their needs. When the CoRS is implemented as a dialog agent, user and recommender interact by exchanging text messages. However, there is little evidence on how effective the interaction is when the CoRS is implemented through a Social Humanoid Robot. In this paper, we evaluate the possibility of introducing an interface based on a Social Humanoid Robot in ConveRSE, a domain-independent framework for the development of Conversational Recommender Systems. The novel interface will be compared against the existing chatbot-based one. The objective is to discover whether the framework can adapt to the new interface without worsening user experience and accuracy. We carried out a preliminary study, which involved 20 subjects. Results proved that, even though there are differences in how users approach the system using the two interfaces, there is no significant difference in its performance.

In this work we propose Ask Me Any Rating (AMAR), a novel content-based recommender system based ... more In this work we propose Ask Me Any Rating (AMAR), a novel content-based recommender system based on deep neural networks which is able to produce top-N recommendations leveraging user and item embeddings which are learnt from textual information describing the items. A comprehensive experimental evaluation conducted on stateof-the-art datasets showed a significant improvement over all the baselines taken into account.

Constraint-based recommenders support customers in identifying relevant items from complex item a... more Constraint-based recommenders support customers in identifying relevant items from complex item assortments. In this paper we present WeeVis, a constraint-based environment that can be applied in different scenarios in the e-government domain. WeeVis supports collaborative knowledge acquisition for recommender applications in a MediaWiki-based context. This paper shows how Wiki pages can be extended with recommender applications and how the environment uses intelligent mechanisms to support users in identifying the optimal solutions to their needs. An evaluation shows a performance overview with different knowledge bases.

Do You Feel Blue? Detection of Negative Feeling from Social Media

AI*IA 2017 Advances in Artificial Intelligence, 2017

The blue feeling is the sensation which affects people when they feel down, depressed, sad and mo... more The blue feeling is the sensation which affects people when they feel down, depressed, sad and more generally when they are in a bad feeling state. In some cases, it is a recurring situation in their everyday life and it can be the first symptom of more complex psychological diseases such as depression. In the last decade, as consequence of the quick increase of detected cases of depression in children and teenagers, it has become very important to find strategies for a timely detection of this pathology. In this work, we describe a model that can support the detection task, by identifying some warning scenarios of blue feeling. The proposed architecture is composed by modules focused on different aspects that characterize the scenario: changes in heart rate, reduction of sleep, reduction of activities performed, increases of use of negative phrases and words. In particular, in this paper, we describe the approach adopted to analyze users posts on social media networks (SMNs) by using natural language processing techniques. The proposed approach is evaluated through an experimental session over a dataset of Facebook posts. The results show good performance in the detection of negative feeling.

Proceedings of the Joint Workshop on Interfaces and Human Decision Making for Recommender Systems, IntRS 2015, Vienna, Austria, September 19, 2015

Proceedings of the 5th Joint Workshop on Interfaces and Human Decision Making for Recommender Systems, IntRS 2018, co-located with ACM Conference on Recommender Systems (RecSys 2018), Vancouver, Canada, October 7, 2018

The popularity of social robots is steadily increasing, mainly due to the interesting impact they... more The popularity of social robots is steadily increasing, mainly due to the interesting impact they have in several application domains. In this paper, we propose the use of Pepper Robot as an interface of a recommender system for tourism. In particular, we used the robot to interact with the users and to provide them with personalized recommendations about hotels, restaurants, and points of interest in the area. The personalization mechanism encoded in the social robot relies on soft biometrics traits automatically recognized by the robot, as age and gender, user interests and personal facets. All these data are used to feed a neural network that returns as output the most suitable recommendations for the target user. To evaluate the effectiveness of the interaction driven by a social robot, we carried out a user study whose goal was to evaluate: (1) how the robot affects the perceived accuracy of the recommendations; (2) how the user experience and the engagement vary by interacting...

Interfaces and Human Decision Making for Recommender Systems

Fourteenth ACM Conference on Recommender Systems, 2020

As an interactive intelligent system, recommender systems are developed to give recommendations t... more As an interactive intelligent system, recommender systems are developed to give recommendations that match users’ preferences. Since the emergence of recommender systems, a large majority of research focuses on objective accuracy criteria and less attention has been paid to how users interact with the system and the efficacy of interface designs from users’ perspectives. The field has reached a point where it is ready to look beyond algorithms, into users’ interactions, decision making processes, and overall experience. The series of workshops on Interfaces and Human Decision Making for Recommender Systems focuses on the ”human side” of recommender systems. The goal of the research stream featured at the workshop is to improve users’ overall experience with recommender systems by integrating different theories of human decision making into the construction of recommender systems and exploring better interfaces for recommender systems. In this summary, we introduce 7th Joint Workshop on Interfaces and Human Decision Making for Recommender Systems at RecSys’20, review its history, and discuss most important topics considered at the workshop.

Towards Emotion-aware Recommender Systems: an Affective Coherence Model based on Emotion-driven Behaviors

Expert Systems with Applications, 2021

Abstract Decision making is the cognitive process of identifying and choosing alternatives based ... more Abstract Decision making is the cognitive process of identifying and choosing alternatives based on preferences, beliefs, and degree of importance given by the decision maker to objects or actions. For instance, choosing which movie to watch is a simple, small-sized decision-making process. Recommender systems help people to make this kind of choices, usually by computing a short list of suggestions that reduces the space of possible options. These systems are strongly based on the knowledge of user preferences but, in order to fully support people, they should be grounded on a holistic view of the user behavior, that includes also how emotions, mood, and personality traits influence her choosing patterns. In this work, we investigate how to include emotional aspects in the recommendation process. We suggest that the affective state of the user, defined by a set of emotions (e.g., joy, surprise), constitutes part of choosing situation that should be taken into account when modeling user preferences. The main contribution of the paper is a general emotion-aware computational model based on affective user profiles in which each preference, such as a 5-star rating on a movie, is associated with the affective state felt by the user at the time when that preference was collected. The model estimates whether an unseen item is suitable for the current affective state of the user, by computing an affective coherence score that takes into account both the affective user profile and not-affective item features. The approach has been implemented into an Emotion-aware Music Recommender System, whose effectiveness has been assessed by performing in-vitro experiments on two benchmark datasets. The main outcome is that our system showed improved accuracy of recommendations compared to baselines which include no affective information in the recommendation model.

IEEE Access, 2020

Recommender systems (RSs) are systems that produce individualized recommendations as output or dr... more Recommender systems (RSs) are systems that produce individualized recommendations as output or drive the user in a personalized way to interesting or useful objects in a space of possible options. Recently, RSs emerged as an effective support for decision making. However, when people make decisions, they usually take into account different and often conflicting information such as preferences, long-term goals, context, and their current condition. This complexity is often ignored by RSs. In order to provide an effective decision-making support, a RS should be ''holistic'', i.e., it should rely on a complete representation of the user, encoding heterogeneous user features (such as personal interests, psychological traits, health data, social connections) that may come from multiple data sources. However, to obtain such holistic recommendations some steps are necessary: first, we need to identify the goal of the decision-making process; then, we have to exploit common-sense and domain knowledge to provide the user with the most suitable suggestions that best fit the recommendation scenario. In this article, we present a methodological framework that can drive researchers and developers during the design process of this kind of ''holistic'' RS. We also provide evidence of the framework validity by presenting the design process and the evaluation of a food RS based on holistic principles.

Information Systems, 2017

HealthNet (HN) is a social network that brings together patients with similar health conditions. ... more HealthNet (HN) is a social network that brings together patients with similar health conditions. HN helps users in finding a solution to their health problems by suggesting doctors and health facilities that best fit the patient profile. Indeed, the core component of HN is a recommender system that suggests patients similar to the target user and supports the choice of the doctor and the hospital for a specific condition. The recommendation algorithm first computes similarities among patients, and then generates a ranked list of doctors and hospitals for a given patient profile by exploiting health data shared by the community. The HN typical user can find the most similar patients, can look how they treated their diseases, and can receive suggestions for solving her condition. In order to facilitate the interaction with the system and improve the recommendation step, the patient can express her health status by a natural-language sentence. The system analyzes the sentence and identifies the most relevant medical area (e.g., orthopedics, neurology, allergology, etc.) for that specific case, and uses this information for the recommendation task. Currently HN is in alpha version and only for Italian users, but in the future we want to extend the platform to other languages. We carried out both an in-vitro experimental evaluation to assess the effectiveness of the module for analyzing natural language descriptions provided by users as well as the recommender system to suggest the right

Information Sciences, 2017

Cross-lingual data linking is the problem of establishing links between resources, such as places... more Cross-lingual data linking is the problem of establishing links between resources, such as places, services, or movies, which are described in different languages. In cross-lingual data linking it is often the case that very short descriptions have to be matched, which makes the problem even more challenging. This work presents a method named TRanslation-based Explicit Semantic Analysis (tr-esa) to represent and match short textual descriptions available in different languages. tr-esa translates short descriptions in any given language into a pivot language by exploiting a machine translation tool. Then, it generates a Wikipedia-based representation of the translated text by using the Explicit Semantic Analysis technique. The resulting representations are used to match short descriptions in different languages. The method is incorporated in CroSeR (Cross-lingual Service Retrieval), an interactive data linking tool that recommends potential matches to users. We compared results coming from an in-vitro evaluation on a gold standard consisting of five datasets in different languages, with an in-vivo experiment that involved human experts supported by CroSeR. The in-vivo evaluation confirmed the results of the in-vitro evaluation and the overall effectiveness of the proposed method.

Emotion Detection Techniques for the Evaluation of Serendipitous Recommendations

Human–Computer Interaction Series, 2016

Recommender systems analyze a user’s past behavior, build a user profile that stores information ... more Recommender systems analyze a user’s past behavior, build a user profile that stores information about her interests, maybe find others who have a similar profile, and use that information to find potentially interesting items. The main limitation of this approach is that provided recommendations are accurate, because they match the user profile, but not useful as they fall within the existing range of user interests. This drawback is known as overspecialization. New methods are being developed to compute serendipitous recommendations, i.e. unexpected suggestions that stimulate the user curiosity toward potentially interesting items she might not have otherwise discovered. The evaluation of those methods is not simple: there is a level of emotional response associated with serendipitous recommendations that is difficult to measure. In this chapter, we discuss the role of emotions in recommender systems research, with focus on their exploitation as implicit feedback on suggested items. Furthermore, we describe a user study which assesses both the acceptance and the perception of serendipitous recommendations, through the administration of questionnaires and the analysis of users’ emotions. Facial expressions of users receiving recommendations are analyzed to evaluate whether they convey a mixture of emotions that helps to measure the perception of serendipity of recommendations. The results showed that positive emotions such as happiness and surprise are associated with serendipitous suggestions.

Proceedings of the Second Italian Conference on Computational Linguistics CLiC-it 2015

Negation for Document Re-ranking in Ad-hoc Retrieval