-
Rethinking negative sampling in content-based news recommendation
Authors:
Miguel Ângelo Rebelo,
João Vinagre,
Ivo Pereira,
Álvaro Figueira
Abstract:
News recommender systems are hindered by the brief lifespan of articles, as they undergo rapid relevance decay. Recent studies have demonstrated the potential of content-based neural techniques in tackling this problem. However, these models often involve complex neural architectures and often lack consideration for negative examples. In this study, we posit that the careful sampling of negative e…
▽ More
News recommender systems are hindered by the brief lifespan of articles, as they undergo rapid relevance decay. Recent studies have demonstrated the potential of content-based neural techniques in tackling this problem. However, these models often involve complex neural architectures and often lack consideration for negative examples. In this study, we posit that the careful sampling of negative examples has a big impact on the model's outcome. We devise a negative sampling technique that not only improves the accuracy of the model but also facilitates the decentralization of the recommendation system. The experimental results obtained using the MIND dataset demonstrate that the accuracy of the method under consideration can compete with that of State-of-the-Art models. The utilization of the sampling technique is essential in reducing model complexity and accelerating the training process, while maintaining a high level of accuracy. Finally, we discuss how decentralized models can help improve privacy and scalability.
△ Less
Submitted 13 November, 2024;
originally announced November 2024.
-
Towards the automation of book typesetting
Authors:
Sérgio M. Rebelo,
Tiago Martins,
Diogo Ferreira,
Artur Rebelo
Abstract:
This paper proposes a generative approach for the automatic typesetting of books in desktop publishing. The presented system consists in a computer script that operates inside a widely used design software tool and implements a generative process based on several typographic rules, styles and principles which have been identified in the literature. The performance of the proposed system is tested…
▽ More
This paper proposes a generative approach for the automatic typesetting of books in desktop publishing. The presented system consists in a computer script that operates inside a widely used design software tool and implements a generative process based on several typographic rules, styles and principles which have been identified in the literature. The performance of the proposed system is tested through an experiment which included the evaluation of its outputs with people. The results reveal the ability of the system to consistently create varied book designs from the same input content as well as visually coherent book designs with different contents while complying with fundamental typographic principles.
△ Less
Submitted 22 October, 2023;
originally announced October 2023.
-
Using Computational Approaches in Visual Identity Design: A Visual Identity for the Design and Multimedia Courses of Faculty of Sciences and Technology of University of Coimbra
Authors:
Sérgio M. Rebelo,
Tiago Martins,
Artur Rebelo,
João Bicker,
Penousal Machado
Abstract:
Computational approaches are beginning to be used to design dynamic visual identities fuelled by data and generative processes. In this work, we explore these computational approaches in order to generate a visual identity that creates bespoke letterings and images. We achieve this developing a generative design system that automatically assembles black and white visual modules. This system genera…
▽ More
Computational approaches are beginning to be used to design dynamic visual identities fuelled by data and generative processes. In this work, we explore these computational approaches in order to generate a visual identity that creates bespoke letterings and images. We achieve this developing a generative design system that automatically assembles black and white visual modules. This system generates designs performing two main methods: (i) Assisted generation; and (ii) Automatic generation. Assisted generation method produces outputs wherein the placement of modules is determined by a configuration file previous defined. On the other hand, the Automatic generation method produces outputs wherein the modules are assembled to depict an input image. This system speeds up the process of design and deployment of one visual identity design as well as it generates outputs visual coherent among them. In this paper, we compressively describe this system and its achievements.
△ Less
Submitted 7 September, 2022;
originally announced September 2022.
-
Dimensional emotion recognition using visual and textual cues
Authors:
Pedro M. Ferreira,
Diogo Pernes,
Kelwin Fernandes,
Ana Rebelo,
Jaime S. Cardoso
Abstract:
This paper addresses the problem of automatic emotion recognition in the scope of the One-Minute Gradual-Emotional Behavior challenge (OMG-Emotion challenge). The underlying objective of the challenge is the automatic estimation of emotion expressions in the two-dimensional emotion representation space (i.e., arousal and valence). The adopted methodology is a weighted ensemble of several models fr…
▽ More
This paper addresses the problem of automatic emotion recognition in the scope of the One-Minute Gradual-Emotional Behavior challenge (OMG-Emotion challenge). The underlying objective of the challenge is the automatic estimation of emotion expressions in the two-dimensional emotion representation space (i.e., arousal and valence). The adopted methodology is a weighted ensemble of several models from both video and text modalities. For video-based recognition, two different types of visual cues (i.e., face and facial landmarks) were considered to feed a multi-input deep neural network. Regarding the text modality, a sequential model based on a simple recurrent architecture was implemented. In addition, we also introduce a model based on high-level features in order to embed domain knowledge in the learning process. Experimental results on the OMG-Emotion validation set demonstrate the effectiveness of the implemented ensemble model as it clearly outperforms the current baseline methods.
△ Less
Submitted 3 May, 2018;
originally announced May 2018.
-
Urgency/Emergency Health Processes' Modelling: A Case Study
Authors:
Carlos Santos,
Armindo Rebelo,
Carla Ferreira,
José Tribolet
Abstract:
The growing complexity and sophistication of the organizational information systems, and hospital ones particularly, render difficult their comprehension and, consequently, the implementation of control mechanisms that may assure, at all times, the auditability of the above mentioned systems, without having to use models. This paper, framed in a wider investigation, aims to describe the applicatio…
▽ More
The growing complexity and sophistication of the organizational information systems, and hospital ones particularly, render difficult their comprehension and, consequently, the implementation of control mechanisms that may assure, at all times, the auditability of the above mentioned systems, without having to use models. This paper, framed in a wider investigation, aims to describe the application of techniques and methodologies, in the sphere of action of Organizational Engineering, in the modelling of business processes developed in the main Operating Theatre of the Coimbra's University Hospital Emergency Service, as a support for the implementation of an information system architecture, using for that purpose the CEO framework, developed and suggested by the Centre for Organizational Engineering (CEO), based on the UML language.
△ Less
Submitted 10 April, 2012; v1 submitted 7 December, 2011;
originally announced December 2011.
-
Identification of the Risk Related to a Process on Hospital Emergency Service: a Case Study
Authors:
Carlos Santos,
Armindo Rebelo,
Carla Ferreira,
José Tribolet
Abstract:
This paper, framed in a vast investigation, describes the application of techniques and methodologies in Organizational Engineering connected to the associated risk to the processes developed in an Emergency Service of an important Portuguese Hospital. The transactions performed in an emergency service and the consequent risk identification (negative behaviour associated to those transactions) is…
▽ More
This paper, framed in a vast investigation, describes the application of techniques and methodologies in Organizational Engineering connected to the associated risk to the processes developed in an Emergency Service of an important Portuguese Hospital. The transactions performed in an emergency service and the consequent risk identification (negative behaviour associated to those transactions) is done based on static and dynamic models, developed during the business modelling. Any non-trivial system is better portrayed trough a small number of reasonably independent models. From this point of view it is important to look at the systems from a "micro" perspective, which allows us to analyse the system at the transaction level. All processes have some associated risk (inherent risk). Its identification will be decisive for future analysis and for the consequent decision over the need, or not, to study internal control mechanisms. This decision will depend on the risk level that the organization considers acceptable.
△ Less
Submitted 10 April, 2012; v1 submitted 7 December, 2011;
originally announced December 2011.
