The limitivism philosophy holds that an accurate connectionist account can only approximate good symbolic descriptions within certain limit.Grounding symbolic structure onto the vector space has been researched in the literature but precise solution has not yet emerged.Here, we present a geometric method that embodies symbolic tree structures precisely onto learned vector representation.This method turns vector embedding of symbols into nested sets of -spheres (spheres in a higher dimensional space), with two desirable properties: (1) each vector embedding is well preserved by the central point of an -sphere; (2) symbolic tree structures are precisely encoded by inclusion relations among -spheres.This unified representation bridges the gap between Deep Learning and symbolic structural knowledge.Significant experiment results are obtained by embodying a large hypernym trees word-sense tree onto GloVE word embeddings of tree nodes.Our geometric method shows a new way to completely resolve the antagonism between connectionism and symbolicism.
As a key element of user experience, communicating expectations to a user through visual elements instead of natural language can produce a more intuitive interface as users interact with a system.While such visual languages are developed by domain experts for specific purposes, these languages can also grow, change, and evolve within a community of users in the same manner as natural language.Video games are one area where communication with the user directly affects the enjoyability, usability, and accessibility of the system.While work has been done to create and leverage visual languages in order to gamify learning or improve accessibility, this research focuses on the creation and evolution of these visual languages by external experts.This exploration of a "crowd-sourced" context-aware visual language through the lens of a structural linguistic framework examines a system of indicators that has evolved over time, created by the very users of the language, to communicate the expectations and necessary actions to complete a task with other members of the community.Extending the analysis and design of visual languages to account for linguistic theory affords designers new tools and approaches in their own work, as too often disciplinary experts can be restricted by conventional understanding, "best practices", and what can be considered a "legitimate" object of study.In this paper, we examine a recent and widely popular visual language of indicators used in the Super Mario Maker games, and show how this use of indicators is central to making a usable text, a playable level, creating a relationship between player and designer that foregrounds the human elements in creating a visual language to assist users in a task.Or, in the case of "Troll" levels, prevent the user from doing so.
Chinese characters are remarkable for the form of art, as the Chinese Calligraphy Painting.However, it is difficult for the visually impaired and people who are unfamiliar with Chinese to experience the beauty of the Chinese characters.In this study, the Sonification scheme, Im2Ms, is proposed to extract the melody between the lines, i.e., the lines in strokes.In Im2Ms, the two-dimensional spatial image information is transformed into the temporal music acoustics domain based on artistic conception and human perception among space, color and sound interaction.Therefore, the Sonification of Chinese Calligraphy Painting not only provide a free access for the visually impaired and people who are unfamiliar with Chinese to appreciation but also enrich the state of mind and imagery in the delivery process.Thus, an immersive appreciation environment of Chinese Calligraphy Painting can be further developed.
This paper deals with the first hypotheses of elevations and of the architectural analysis of San Ildefonso’s Baths on the basis of unpublished data offered by the Book of written descriptive records (apeos) of 1542 in the Chapter on Seville Cathedral. Our own transformation methodology has been applied to this hitherto largely unknown book in order to attain and encompass graphic representation from the literary description. After giving a brief history of the Baths, the objective is to ascertain its location, by drawing hypothetical floor plans and elevations and analysing its typology, dimensions, spaces, and building elements. This research has shown that it was one of the most important bathing complexes in the historical centre of Seville, at least in the 13th Century, although it later became obsolete and was demolished in the 18th Century.
High-quality Artificial intelligence (AI) software in different domains, like image recognition, has been widely emerged in people’s daily life. They are built on machine learning models to implement intelligent features. However, the current research on image recognition software rarely discusses test questions, clear quality requirements, and evaluation methods. The quality of image recognition applications becomes more and more prominent. A three-dimensional(3D) classification decision table can help users to conduct classification-based test requirement analysis and modeling for any given mobile apps powered with AI functions in detection, classification, and prediction. This paper presents a case study of a realistic image recognition application called Calorie Mama using manual testing and automation testing with a 3D decision table. The study results indicate the proposed method is feasible and effective in quality evaluation.
The paper analyses the addition of project management features of an LMS, a topic that LMSs and collaboration platforms have entirely ignored.Many project management-like platforms are available today, especially riding the media wave of agile methodologies.The question is not surprising in itself, as managing time, costs, and resources linked to the discipline of Project Management, is historically a problematic issue for the IT world from a cultural, technical, and organizational point of view.The problem becomes even more complex if applied to the management of learning projects, where we have time, constraints, deadlines, costs, resources like in any other project.LMSs and, in general, collaboration platforms do not include these features, forcing users to link external platforms or adapt existing features (like simple to-do lists) to more complex Project Management tasks.In this work, we will present the tests conducted in a collaborative platform based on the metaphor of virtual learning communities.These tests were on a component developed explicitly for managing projects, activities, and resources, integrated inside the LMS with all the other services (blog, forum, file sharing, calendar, reminders, etc.).The introduction of this new component within the system addresses the need to manage collaborative activities between learners, providing a tool for managing and controlling the progress of the activities assigned to the various community members..
This paper describes an approach for supporting automatic satire detection through effective deep learning (DL) architecture that has been shown to be useful for addressing sarcasm/irony detection problems.We both trained and tested the system exploiting articles derived from two important satiric blogs, Lercio and IlFattoQuotidiano, and significant Italian newspapers.
The COVID-19 pandemic has caused disruption across the globe and put pressure on healthcare systems.In order to limit the use of hospital resources, the use of home care and telehealth has been very important to minimize direct human intervention in monitoring patients.The purpose of this work is to present YouCare: a cross-platform application that allows the collection of medical data on the health status of the user in order to allow physicians to efficiently monitor the status of the patient.As an important feature, it includes functions to monitor the general situation through statistics and interactions with the other users of the application.This might make the isolation period less stressful while exchanging current COVID experiences.The use of the application has been experimented with a usability test, obtaining positive feedback from the users.We also report other similar applications that have been developed and used in different parts of the world.
Self-localization of mobile objects is a fundamental requirement for autonomy. Mobile objects can be for example a mobile service robot, a motorized wheelchair, a mobile cart for transporting tasks or similar. Self-localization represents as well a necessary feature to develop systems able to perform autonomous movements such as navigation tasks. Self-localization is based upon reliable information coming from sensor devices situated on the mobile objects. There are many sensors available for that purpose. The early devices for positioning are rotary encoders. If the encoders are connected to wheels or legs movement actuators, relative movements of the mobile object during its path [3] can be measured. Then, mobile object positioning can be obtained with dead-reckoning approaches. Dead reckoning [3] is still widely used for mobile robot positioning estimation. It is also true that dead-reckoning is quite unreliable for long navigation tasks, because of accumulated error problems.
In the biological field, having a visual and interactive representation of data is useful, particularly when there is a need to investigate a large amount of multilevel data. It is advantageous to communicate this knowledge intuitively because it helps the users to perceive the dynamic structure in which the correct connections are present and can be extrapolated. In this work, we propose a human-interaction system to view similarity data based on the functions of the Gene Ontology (Cellular Component, Molecular Function, and Biological Process) of the proteins/genes for Alzheimer disease and Parkinson disease. The similarity data was built with the Lin andWangmeasures for all three areas of Gene Ontology. We clustered data with the K-means algorithm in order to demonstrate how information derived from data can only be partial when using traditional display methods. Then, we have suggested a dynamic and interactive view based on SigmaJS with the aim of allowing customization in the interactive mode of the analysis workflow by users. To this aim, we have developed a first prototype to obtained a more immediate visualization to capture the most relevant information within the three vocabularies of Gene Ontology. This facilitates the creation of an omic view and the ability to perform a multilevel analysis with more details which is much more valuable for the understanding of knowledge by the end users.
In this paper, we suggest SENECA, a tool that attempts to assist students who follow remote classes in maintaining/capturing attention, allowing them to focus on context-driven learning. Distance education has a number of disadvantages, including a lack of physical interaction between students and teachers, emotional and motivational isolation as a result of this strategy, and a reduction in active engagement. All of these things have an impact on student learning abilities. The largest distractions at home are considered among these disadvantages of distant education, particularly for subjects with low awareness. These distractions cause a movement of the student’s attention from the current lesson to disturbing events. For this reason, there is a need to experiment with new solutions also linked to Information Technology (IT) to improve the focused learning during distance education. Our tool’s technical idea is to create a real-time summary of the topic treated by the teacher. The system captures the text every five minutes, generates outlines, and browses them to eliminate repetitive portions after each survey. We looked at two different sorts of filters, semantic and summary, to see if the first could distinguish between topics and the second could evaluate the topic’s highlights. Natural Language Processing algorithms are used to extract categories and keywords from the general generated summary. The latter will emphasize the most important points of the speech, while the keywords will be utilized to extract the candidate literature about the discussed topics.
The behavior of the NSLPK authentication protocol is visualized using SMGA so that human users can visually perceive non-trivial characteristics of the protocol by observing graphical animations.These characteristics could be used as lemmas to formally verify that the protocol enjoys desired properties.We first carefully make a state picture design for the NSLPK protocol to produce good graphical animations with SMGA and then find out non-trivial characteristics of the protocol by observing its graphical animations.Finally, we also confirm the correctness of the guessed characteristics using model checking.The work demonstrates that SMGA can be applied to the wider class of systems/protocols, authentication protocols in particular.
Bike-sharing is adopted as a valid option replacing traditional public transports since they are eco-friendly, prevent traffic congestions, reduce any possible risk of social contacts which happen mostly on public means. However, some problems may occur such as the irregular distribution of bikes on related stations/racks/areas, and the difficulty of knowing in advance what the rack status will be like, or predicting if there will be bikes available in a specific bike-station at a certain time of the day, or if there will be a free slot to leave the rented bike. Thus, providing predictions can be useful to improve the service quality, especially in those cases where bike racks are used for e-bikes, which need to be recharged. This paper compares the state-of-the-art techniques to predict the number of available bikes and free bike-slots in bike-sharing stations (i.e., bike racks). To this end, a set of features and predictive models were compared to identify the best models and predictors for short-term predictions, namely of 15, 30, 45, and 60 minutes. The study has demonstrated that deep learning and in particular Bidirectional Long Short-Term Memory networks (Bi-LSTM) offers a robust approach for the implementation of reliable and fast predictions of available bikes, even with a limited amount of historical data. This paper has also reported an analysis of feature relevance based on SHAP that demonstrated the validity of the model for different cluster behaviours. Both solution and its validation were derived by using data collected in bike-stations in the cities of Siena and Pisa (Italy), in the context of Sii-Mobility National Research Project on Mobility and Transport and Snap4City Smart City IoT infrastructure.
The work describes a module that has been implemented for being included in a social humanoid robot architecture, in particular a storyteller robot, named NarRob.This module gives a humanoid robot the capability of mimicking and acquiring the motion of a human user in real-time.This allows the robot to increase the population of his dataset of gestures.The module relies on a Kinect based acquisition setup.The gestures are acquired by observing the typical gesture displayed by humans.The movements are then annotated by several evaluators according to their particular meaning, and they are organized considering a specific typology in the knowledge base of the robot.The properly annotated gestures are then used to enrich the narration of the stories.During the narration, the robot semantically analyses the textual content of the story in order to detect meaningful terms in the sentences and emotions that can be expressed.This analysis drives the choice of the gesture that accompanies the sentences when the story is read.
The wide spreading of the Internet leads to the born of a whole interconnected world. Among all these devices, smart voice assistants are gaining particular attention thanks to their ease of use, allowing users to comfortably deploy commands for controlling other devices. The simplicity of use of voice assistants allowed non-expert to interact with complex systems, leading to that category of users with limited knowledge, to interact with s without being perfectly aware of the risks they are exposed to. For example, common network monitoring systems are so useful as they are complex to use for non-expert users. This paper presents a framework for intrusion detection specifically designed to be used by any category of users, using visual interfaces for simplifying the user interaction with the framework, allowing him/her to properly configure and run an Intrusion Detection System (IDS). The implementation of voice assistants as a communication channel will further improve the overall user experience.
In this paper we describe a falls detection and classification algorithm for discriminating falls from daily life activities using a MEMS accelerometer.The algorithm is based on a shallow Neural Network with three hidden layers, used as fall/non fally classifier, trained with daily life activities features and fall features.The novelty of this algorithm is that synthetic falls are generated as multivariate random Gaussian features, so only real daily life features must be collected during some day of normal living.Moreover, the features related to synthetic fall events are generated as complement of normal features.First of all, the features acquired during daily life are clustered by Principal Component Analysis and no Fall activities shall be recorded.The complement set of the normal features is found and used as a mask for Monte Carlo generation of synthetic fall.The two feature sets, namely the features recorded from daily life activities and those artificially generated are used to train the Neural Network.This approach is suitable for a practical utilization of a Neural Network based fall detection characterized by high Recall-Precision rate.
The effective design and delivery of assessments in a wide variety of evolving educational environments remains a challenging problem. Proposals have included the use of learning dashboards, peer learning environments, and grading support systems; these embrace visualisations to summarise and communicate results. In an on-going project, the investigation of graph based visualisation models for assessment design and delivery has yielded promising results. Here, an alternative graph foundation, a two-weighted hypergraph, is considered to represent the assessment material (e.g., questions) and their explicit mapping to one or more learning objective topics. The visualisation approach considers the hypergraph as a collection of levels; the content of these levels can be customized and presented according to user preferences. A case study on generating hypergraph models using commonly available assessment data and a flexible visualisation approach using historical data from an introductory programming course is presented
Several approaches have emerged to support the development of mobile web applications (apps).Front-end frameworks (FeF) have emerged to support developers in the construction of responsive mobile web apps.However, these frameworks do not provide resources to handle easily variables of the context of use and to deal with different modalities of interaction.Considering this gap, we proposed the HyMobWeb, an approach that assists developers in working with these aspects.By grasping the popularity of the FeFs, HyMobWeb proposes a flexible and reusable approach based on FeF structure.It works with a hybrid approach that treats the adaptation in two phases.The static one, performed in the development time, allows developers to implement the resources of adaptation.The dynamic one performs the adaptation during the run-time.In this article, we present the HyMobWeb dynamic adaptation approach.As end-users are the receivers of the dynamic adaptation our evaluation was carried out in the perspective of this audience.The results showed that the adaptations regarding the context of use were the ones that presented more impact on the user experience.We concluded that the HyMobWeb dynamic adaptation provides ways to enhance user interaction in mobile web apps when compared with RWD resources.
